ArticlePDF Available

Abstract and Figures

The complexity of the world oil market has increased dramatically in recent years and new approaches are needed to understand, model, and forecast oil prices today. In addition to the commencement of the financialization era in oil markets, there have been structural changes in the global oil market. Financial instruments are communicating information about future conditions much more rapidly than in the past. Prices from long and short duration contracts have started moving more together. Sudden supply and demand adjustments, such as the financial crisis of 2008-2009, faster Chinese economic growth, the Libyan uprising, the Iranian nuclear standstill or the deepwater horizon oil spill, change expectations and current prices. The daily Brent spot price fluctuated between $30 and above $140 per barrel since the beginning of 2004. Both fundamental and financial explanations have been offered as explanatory factors. This paper selectively reviews the voluminous literature on oil price determinants since the early 1970s. It concludes that most researchers attribute the long-run oil price path to fundamental factors such as economic growth, resource depletion, technical advancements in both oil supply and demand, and the market organization of major oil petroleum exporting countries (OPEC). Short-run price movements are more difficult to explain. Many researchers attribute short-run price movements to fundamental supply and demand factors in a market with very little quantity response to price changes. Nevertheless, there appears to be some evidence of occasional financial bubbles particularly in months leading up to the financial collapse in 2008. These conflicting stories will not be properly integrated without a meeting of the minds between financial and energy economists.
Content may be subject to copyright.
Oil Markets and Price Movements: A Survey of Determinants
Hillard Huntingtona, Zhuo Huangb, Saud M. Al-Fattahc, Michael Gucwaa, and Ali Nouria
aEnergy Modeling Forum, Huang Engineering Center, Stanford University, 475 Via Ortega,
Stanford, CA 94305-4121.
bPeking University, Beijing 100871, China.
cKing Abdullah Petroleum Studies and Research Center, P.O. Box 8349, Dhahran 31311, Saudi
Arabia; Email:
The complexity of the world oil market has increased dramatically in recent years and new
approaches are needed to understand, model, and forecast oil prices today. In addition to the
commencement of the financialization era in oil markets, there have been structural changes in the
global oil market. Financial instruments are communicating information about future conditions
much more rapidly than in the past. Prices from long and short duration contracts have started
moving more together. Sudden supply and demand adjustments, such as the financial crisis of 2008-
2009, faster Chinese economic growth, the Libyan uprising, the Iranian nuclear standstill or the
deepwater horizon oil spill, change expectations and current prices.
The daily Brent spot price fluctuated between $30 and above $140 per barrel since the
beginning of 2004. Both fundamental and financial explanations have been offered as explanatory
factors. This paper selectively reviews the voluminous literature on oil price determinants since the
early 1970s. It concludes that most researchers attribute the long-run oil price path to fundamental
factors such as economic growth, resource depletion, technical advancements in both oil supply and
demand, and the market organization of major oil petroleum exporting countries (OPEC). Short-run
price movements are more difficult to explain. Many researchers attribute short-run price
movements to fundamental supply and demand factors in a market with very little quantity response
to price changes. Nevertheless, there appears to be some evidence of occasional financial bubbles
particularly in months leading up to the financial collapse in 2008. These conflicting stories will not
be properly integrated without a meeting of the minds between financial and energy economists.
Keywords: Oil prices, supply and demand analysis, financial markets.
Abstract ........................................................................................................................................... 1
1. Introduction .................................................................................................................................... 3
2. Crude Oil Price Behavior ............................................................................................................... 4
2.1 Price Persistence and Volatility................................................................................................... 5
2.2 Drivers of Long-Run Crude Oil Price Paths ............................................................................. 7
2.3 The Current Conventional Wisdom ............................................................................................ 7
3. Oil Demand and Responses .......................................................................................................... 9
3.1 Growth and Industrialization ................................................................................................... 10
3.2 Oil Demand and Technical Progress ....................................................................................... 12
3.3 Alternative Vehicles and Competitive Fuels............................................................................ 14
3.4 Demand Response to Oil Prices .............................................................................................. 14
4. Oil Supply Availability and Costs ................................................................................................ 16
4.1 Resources and Geological Availability ..................................................................................... 16
4.2 Resource Costs .......................................................................................................................... 18
4.3 Extraction of a Depletable Resource ....................................................................................... 20
4.4 Oil Supply from Competitive Regions ..................................................................................... 22
5. OPEC Behavior ............................................................................................................................ 23
5.1 Conceptual Approaches for Understanding OPEC Behavior ............................................... 23
5.2 Exporter Behavioral Strategy .................................................................................................... 25
5.3 Empirical Studies of OPEC Behavior ..................................................................................... 26
5.4 National Oil Companies ........................................................................................................... 27
6. Stylized Facts about the Short-Run Dynamics of Oil Prices .................................................... 29
6.1 Higher Volatility beyond the Explanation of Fundamental Factors ..................................... 29
6.2 Closer Link with Macroeconomic Variables and the Prices of Other Commodities ........... 29
6.3 Flat Forward Curves .................................................................................................................. 31
7. Underlying Factors Affecting Short-Run Oil Prices .................................................................. 33
7.1 Demand and Supply Shocks ..................................................................................................... 33
7.2 Macroeconomic Variables ......................................................................................................... 35
7.3 Financialization and Speculation .............................................................................................. 39
7.4 Financialization as a Double-Edged Sword ............................................................................. 57
8. Expectation and the Linkage between the Short- and Long-Run Price Movements .............. 60
8.1 The Oil Price Explosion of 2004-2008 .................................................................................... 60
8.2 Risk Premium Embedded in the Long-Dated Oil Futures Prices ........................................... 62
9. Conclusions ......................................................................................................................................... 64
Acknowledgement ........................................................................................................................................... 66
References ......................................................................................................................................................... 66
Planning oil strategies requires investments based upon long-run fundamental conditions in the
petroleum market and their impact on future crude oil prices. Frequently, it is sufficient to focus on
long-term factors for explaining these trends. Occasionally however, near-term oil price volatility
creates massive uncertainty about fundamental conditions. Financial cycles can also raise borrowing
costs and reduce investment. In these cases, investors need to understand short-run conditions if
they are to evaluate long-run opportunities appropriately.
This paper reviews the determinants and drivers of oil price movements.1 Emphasis is
placed on selected past contributions that appear most critical for evaluating oil price trends in the
coming decades rather than on an exhaustive review of all papers in this very interesting and lengthy
literature.2 The paper explains the major factors considered by past research but does not try to
identify the many possible areas for future investigation.
The paper begins with identifying longer-run oil price drivers before shifting to short-run
developments. Section 2 is a very brief discussion of crude oil price behavior and volatility. Section
3 shows how the long-run oil price path is related to developments in oil demand particularly in
emerging markets. After discussing resources, costs and oil supplies from regions outside the
Arabian Gulf in section 4, the paper focuses on the conceptual and empirical research geared
towards understanding oil production and pricing decisions of the members of the Organization for
Petroleum Exporting Countries (OPEC) in section 5. The remainder of the paper discusses near-
term price volatility and the interrelationship between financial markets on the one hand and the
short-run physical supply-demand fundamentals on the other. Section 6 emphasizes the stylized
facts about the short-run dynamics of oil prices, section 7 explores a number of underlying factors
affecting short-run oil prices, and section 8 emphasizes the role of expectations in providing an
important linkage between the short- and long-run price movements. In summary, the paper’s main
conclusions are as follows:
1. The underlying drivers for long-run oil price paths identified in the literature are:
Rapid economic growth and the resultant oil demand growth in the emerging economies will
exert pressures on oil prices over the coming decades.
Conventional oil supplies will not grow as rapidly due to depletion in aging fields and/or
political and investment barriers.
Unconventional oil supplies from heavy oil, oil sands, and oil shale will significantly expand
the long-run supply options but uncertainty exists about their long-run finding and
development costs. Investors in these resources will undoubtedly require an oil price path
that exceeds these costs to offset the risks created by very volatile oil prices.
1 This paper builds on an overview of prior research compiled by the KAPSARC and Energy Modeling Forum
(KAPSARC 2011).
2 See Fattouh (2007, 2010a) for other recent valuable surveys of the oil market and Al-Qahtani et al (2008) for a
comprehensive review of various theories and models of OPEC behavior.
Although OPEC is now 50 years old, no integrative or comprehensive theory has been
developed to explain the behavior of its members. This vacuum explains one of the major
challenges in understanding future crude oil price trajectories.
Changes in oil production and in transportation technologies have contributed structural
changes in oil markets and are expected to continue to influence future market dynamics.
Government energy policies influencing both production and consumption often have
dramatic impacts on energy and oil markets. Government policies introduce market
imperfections that frequently are challenging for modelers to represent well. Similar
problems arise when a model tries to incorporate the complex relationship between energy
markets and strategic state behavior.
2. The underlying drivers for short-run oil price dynamics mask many of these long-run
developments. Many of these are related to the interaction between financial and physical
markets for crude oil.
Demand and supply shocks, together with the news and the uncertainty associated with the
shocks, remain the primary drivers underlying the short-run oil price dynamics. The shocks
have been magnified by the very low short term price elasticity of both oil supply and
Macroeconomic variables, such as economic growth, exchange rates and interest rates, have
contributed to oil price fluctuations.
Despite the rapidly increased presence of index and other types of financial investors in the
oil futures markets, the empirical evidence for a speculative bubble is weak and the direct
link between the oil price surge of 2004-2008 and speculation has not empirically been
The financialization phenomenon still has profound impact on crude oil markets. It adds
liquidity to the futures market, especially for the long-dated oil futures contracts, which
facilitates hedging and risk management. It has strengthened the co-movements between
front-end and long-dated futures oil prices.
Under certain circumstance, the expectation can be distorted by market sentiment which
could be either excessively optimistic or pessimistic.
A deeper understanding of the role of expectation and the interaction of short-run and long-
run factors in the formation of oil futures prices is needed. The expectation incorporates
existing market information about supply and demand fundamentals in both the short run
and the long run.
The oil futures prices also embed a risk premium that reflects the aggregate risk preference
of market participants.
Petroleum is an extremely fungible commodity where transportation costs for moving oil between
regions are relatively low. When market forces raise the price of crude oil in a particular region,
those dynamics will cross over into other oil markets and influence the price of oil everywhere. Oil’s
fungibility stands in marked contrast to other basic commodities where product differentiation is
quite evident (Nordhaus 2009), and even natural gas which is relatively expensive to transport over
long distances.
This paper will discuss the crude oil price trend and will not evaluate the factors creating
price differentials between crude types and locations, analyzing spreads between the international
benchmarks such as West Texas Intermediate (WTI) versus Brent, or of different maturities, nor will
it analyze different determinants of crude oil and petroleum product prices. All such play important
roles in explaining short-term price movements. The sulfur content, density and other characteristics
of the crude oil often will create differences between oil prices of different types, particularly with
the tightening of environmental standards on petroleum products and bottlenecks in refining
(Verleger 2009). Over longer periods, the prices of these different crude types will move in tandem
with each other when supply and demand factors shift, as long as refinery adjustments allow crude
types to be easily substituted for each other.
2.1 Price Persistence and Volatility
Crude oil price behavior exhibits several distinguishable patterns as shown in Figure 1, which plots
the annual average US crude oil prices since 1985. These crude oil prices have been adjusted for
inflation and are expressed in 2009 US dollars.
Figure 1 – Annual Average WTI Prices, 1985-2011 (2009 USD per barrel)
Sources: Energy Information Administration; Bureau of Labor Statistics
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011
Price (in 2009 USD per Barrel)
First, oil prices have been very volatile, beginning in 1986 and particularly in the 2000-2009
decade. This pattern is even more pronounced for quarterly or monthly frequencies. The standard
deviation of logarithmic oil price changes is frequently used to measure this volatility. This volatility,
increased in the last several decades compared to the 1950s and 1960s due to the changes in the
structure of the oil market and consequently in oil price determination. Another issue is oil price
levels. When prices suddenly change, the question whether they revert back to their previous trend
or take new path?3
Pindyck (1999) evaluates long-run oil prices over the 1861-1998 period as one continuous
series without any breaks. He concludes that oil prices eventually revert back to a long-run, U-
shaped historical trend, although this adjustment happens over a long period (seven years or more).
When estimated as a moving quadratic function, this trend declines initially before rising over time.
Hamilton (2009b) reached a different conclusion about oil prices in evaluating the 1973-2008
period with quarterly rather than annual data. He finds that oil prices over this period are not mean-
reverting but instead deviated around their level in the previous period. This result implies a “no-
change” rule for estimating future prices; the best guess for next period’s price will be last period’s
Dvir and Rogoff (2009) analyze the 1861-2008 oil price series as three different epochs.
They find periods when both oil price trends and volatilities changed. Oil prices in the 1861-1877
and the 1972-2008 periods were both volatile and persisted at relatively high levels. These periods
were characterized by rapid growth in a major economy (the United States and Asia, respectively)
and by limited spare oil capacity (by US railroad companies and OPEC members, respectively). In
contrast, oil prices displayed much less volatility during the 1877-1972 period. Although this era had
several major demand expansions, the Texas Railroad Commission could ease its control of the
spacing of wells to accommodate any sudden increase in oil demand. Adelman (1964) described the
early 1960s oil market as neither fully monopolistic nor a competitive market where the price would
be set lower by marginal costs. The higher price trends in the first and third periods of Dvir and
Rogoff study and the lower price trend in the second period are consistent with the U-shape trend
found by Pindyck (1999). Considering the two most recent periods, Libecap and Smith (2004)
compared the controlling role of Texas Railroad Commission (1933-1972) in United States and
Saudi Arabia in OPEC (since 1973). Their findings suggest that the efforts of Texas Railroad
Commission were much more centralized than OPEC in the sense that there is no central unit in
OPEC for monitoring and enforcing agreements.
The Dvir-Rogoff findings suggest that, over the last 140 years, prices tend to be highly
persistent and volatile when unexpected and rapid growth in a major economy is accompanied by
uncertain access to additional supplies. These conditions should continue in the future, because
uncertainty about the speed and duration of rapidly growing economies in Asia are occurring
simultaneously with more limited oil supplies and investment outside of the Arabian Gulf. Prices
should remain high and volatile as long as these conditions prevail.
3 Although it can be easier to visualize volatility and persistence with higher-frequency data (e.g., monthly), the Dvir-
Rogoff (2009) study was based upon the annual data in Figure 1. The annual price path in this figure captures price
swings more than monthly or daily fluctuations or even inter-day or intra-day price volatility.
2.2 Drivers of Long-Run Crude Oil Price Paths
There exists unprecedented uncertainty about future oil price movements because they are
influenced in the long run by a number of key factors spanning economics, politics and technology.
The revolution in financial instruments in commodity markets including oil and the more intense
volatility in near-term prices may contribute additional risks for long-run investors in the oil
industry. So long-term investors do not know for certain whether it is the financial market
developments or physical market forces that are driving oil price movements today; they are likely to
remain very uncertain about the true value of the resource they are investing in.
This section explains the major drivers of the long-run oil price path. Long-run price paths
are set by supply and demand conditions in the physical market for crude oil and the various
products derived from it. Supply and demand conditions will change over time, causing oil to be
more or less expensive to find, extract, refine and market. These market changes will also make oil
either more or less costly to use as economies grow. As is the case in most markets, prices adjust to
these changes until markets clear where buyers and sellers reach an agreement on the price.
Reasonable variations in these conditions, however, can lead to a wide band for future oil prices, as
will be discussed in the next section.
2.3 The Current Conventional Wisdom
Oil market analysts frequently focus on the call on OPEC oil” being the difference between
projected world demand and the projected full potential of non-OPEC supply l. OPEC members
take the prevailing price but frequently they will adjust their production to alter their income path
and the logic of the market implies that such adjustments will influence the price as well as the
volume of oil shipped.
Many outlooks by government institutions, consultants and analysts predict future inflation-
adjusted (real) oil prices to grow from current levels. The US Energy Information Administration
(2010), for example, projects that oil prices will fluctuate between $51 to $210 per barrel in constant
2008 dollars by 2035, depending upon whether low-price or high-price conditions prevail. Their
reference projections ($133 in 2030) is compared to OPEC’s (2011) assumption of oil prices
reaching $133 in nominal terms by 2035 but staying in the $85-95 range throughout the 2010-2020
decade. The International Energy Agency (2011) outlook on the other hand assumes prices to reach
$120 in real terms by 2035 in its ‘current policies’ scenario (see appendix K).
Gately (2004, 2007) criticizes price projections by the EIA as being too low. He concludes
that the net present value of OPEC’s income stream in these EIA projections is far inferior to what
they could earn if they restricted their oil exports. Rather than expanding their share of total oil
consumed, OPEC may try to maintain their existing share which, he argues would result in higher
global oil prices.
The world of high-prices embodied in either the EIA or Gately projections adopt
assumptions that provide one reasonable outlook given what is known about supply and demand
conditions, but the discussion below will also offer others. Rapid industrialization, GDP growth and
oil demand growth in many Asian countries, perhaps expanded to other emerging economies as
well, are the principal drivers of these oil price paths. Demand for mobility is likely to grow
dramatically in many emerging economies. This demand growth will be combined with limited oil
supplies due to scarce oil resources and no new discovered giant fields outside of the Arabian Gulf
(although recent and auspicious discoveries in new basins offshore Brazil and West Africa may alter
the calculation). Closing this demand-supply imbalance will require large oil price changes because
supplies and demand are only modestly sensitive to oil prices, even in the long run. Most
projections still see a commanding role for oil over the next three decades whereby transformative
technical progress in vehicles, the emergence of a serious competitor to oil products within the
transportation sector, or tighter climate change policies, will either be slow to penetrate, develop,
accept or implement, respectively
Perceptions today are very much like those that were shaped by world oil market models in
1980 (e.g., see Energy Modeling Forum, 1982). In hindsight, many of these views became badly
outdated by a confluence of offsetting factors: technical progress in oil exploration, world recession
and lower economic growth rates, and technical change in energy use introduced by the rapid oil
price escalation of the 1970s (Huntington, 1994). Although current conditions portend a bleak
future for oil consumers, the world is not necessarily locked into these trends. There exists enough
uncertainty about fundamental supply and demand parameters to support a range of very different
future oil price paths.
Figure 1 presents some major oil price drivers and their effects along two dimensions: the
future crude oil price path and the trend for Arabian Gulf production. Oil prices and Gulf
production in a future year rest at an equilibrium level where the two dashed lines intersect. If new
conditions represented inside one of the four boxes emerge, the price and production will move in
the direction shown in that figure. The box in the upper-right portion of the diagram represents the
prevailing outlook provided by the EIA and other major organizations. Rapidly industrializing Asia
and limited oil resources outside of the Arabian Gulf are two of the more influential factors behind
their projections. Continued economic diversification within the Middle East and other major oil-
exporting nations will contribute to the increased pressure on available oil supplies. The oil price
increases along the vertical axis and production expands along the horizontal axis.
The lower two portions represent a set of alternative conditions that could usher in lower
future oil prices than what many expect today. They include rapid expansion in Iraqi oil production
as well as faster development of renewables and natural gas and reforming local energy prices in the
Gulf. Although Gulf oil production expands under these conditions, the price path is lower.
The lower left box represents conditions where there are stronger international
commitments and coordination in global climate change policy and perhaps energy security
initiatives. Even if oil-consuming nations do not implement these policies, it is likely that investors
will be stimulated by today’s higher prices and the fears of eventual climate and security policies in
the future to bring on new energy sources that will compete with petroleum. The development of
significant substitutes for oil in automobiles like compressed natural gas, diesel from natural gas and
electric vehicles may result from lower prices for these alternatives or from government policies
mandating the switch to these sources. Additionally, some important sources of unconventional oil
production (e.g., Canadian oil sands) appear cost effective at the 2011 prices. These new sources
could restrain future oil price increases if they can be expanded to scale at their current costs. Each
of these important oil-price drivers will be discussed in turn below under the general headings of
demand, non-OPEC production and OPEC production.
Figure 1 – Long-Term Oil Market Factors.
If the world economy continues its rapid growth, oil demand is expected to increase steadily based
upon the estimates of the price and income elasticities. Operating against these trends, however, will
be a set of potentially offsetting factors, some of which outside of the oil market, that could curb
future oil demand growth. At the moment, it is unknown whether these additional effects will be
modest or transformative.
First, gasoline consumption in many emerging economies may not rise as quickly as current
rates as those economies continue to expand. They may be limited not by oil prices and household
budgets but by road congestion and poor urban planning (Schipper et al. 2009). Second, automobile
Gulf Oil Exports
Rapid Asian industrialization
Costly non-Gulf production
Slow development of
alternatives to oil
Greater domestic oil
consumption in the Gulf
Unfavorable geopolitical
Lower oil production in the
Rapid Iraqi expansion
Accelerated development of
renewables and natural gas in
the Gulf
Reforming local energy prices
in the Gulf
Electric vehicles
Compressed natural gas
Diesel from natural gas
Tighter climate policies
Removing gasoline subsidies
in the main consuming
Technical progress in
unconventional oil
Initial Price/Export Equilibrium
Oil Price
producers have placed a high premium on improved fuel efficiency in new vehicles, and these trends
are being reinforced by government mandates. Third, new fuels or mixes are competing seriously
with oil products, as represented by a range of alternatives including gasoline blends with ethanol,
plug-in hybrid vehicles, and natural gas buses and cars. Fourth, petroleum is not the only source for
liquid fuels. More competitive natural gas may encourage gas-to-liquid (GTL) technology that
makes diesel fuel from more abundant natural gas rather than oil. Fifth, many emerging economies,
including some major oil exporting ones, are likely to phase out or eliminate their price subsidies for
fuels. This will thereby curtail their demand growth. Finally, more major industrialized economies
outside the European Union may adopt policy measures that price future carbon emissions and
invest in research and development for new carbon-saving technologies. These policies could
increase the incentives to replace gasoline and other oil products with alternative energy sources.
Global consumption of residual fuel oil has declined from 27% of all petroleum
consumption in the 1970s to 10% in 2010.4 Middle distillates, including jet fuel, have replaced these
declining sales, while world gasoline consumption has grown slightly from 29 to 33%. Much of the
global oil consumption growth of 28 MBD between 1985 and 2010 has occurred outside the OECD
nations, of which the emerging economies accounted for 82 percent of this growth in global
consumption. As a share of world consumption, the emerging nations’ share has grown from 22%
to 41% over this period. The major contributors to this growth experience have been the growth
and industrialization within the world economies, although higher oil prices and technical progress
have offset this trend to some extent.
3.1 Growth and Industrialization
The growth in per capita oil demand is comparable to economic growth in many emerging
economies, especially in rapidly developing Asian nations. Many countries are experiencing rapid
growth in vehicle ownership as incomes, mobility and urbanization rise and an expanding
proportion of the population is earning incomes that allow them to afford private vehicles.
Moreover, as incomes increase, households tend to shift to different types of fuels (cleaner and
more costly) in addition to consuming more of the same fuel (Hosier, 2004).
The vehicle-income relationship appears to follow a Gompertz function shape as shown in
Figure 2, where vehicle ownership grows dramatically during a boom period at lower to middle
income levels before tapering off at higher income levels (Dargay et al. 2007). An important
difference in various studies (e.g., between Dargay et al, 2007, and the US Energy Information
Administration, 2010) is the vehicle saturation point where vehicle ownership per household ceases
to increase. The European saturation point appears sooner than its USA counterpart, reflecting its
greater population density, road congestion and urban transportation infrastructure. Reflecting
similar constraints, the Asian saturation point may even appear sooner than its European
counterpart. Road congestion may be an important limitation, although the Chinese appear
committed to growing a domestic automobile industry and both China and India are developing
plans to build miniature passenger vehicles to reduce the pressure on limited roadways.
4See BP Statistical Review of World Energy 2011
Figure 2Vehicle Ownership and Per Capita Income, 1960-2002.
Source: Dargay et al. (2007).
These vehicle-ownership patterns appear reflected in total oil and energy consumption in
emerging economies, where the income elasticity of oil demand begins to fall as the economy
matures and reaches higher income levels. In studying the energy intensity trends in ten Asian
emerging countries, Galli (1998) finds that income elasticities decrease as incomes increase. In 1973,
all income elasticities exceeded unity and approached almost two in the poorest nation. This result
means that oil demand increases at least as fast as economic growth for all developing countries,
holding constant energy prices and technical progress. In the poorest Asian nations, oil demand
expands twice as fast as economic growth. Medlock and Soligo (2001) have also discovered similar
trends for a larger set of countries and for a more detailed data set disaggregated by sectors. Van
Benthem and Romani (2009) find nonlinear effects for both economic growth and prices for
economic sectors of 24 non-OECD countries. For non-OECD countries with a per capita GDP of
less than $5,000, the energy demand’s elasticity increases with income, but for countries with a GDP
greater than $5,000, the income elasticity decreases; this is consistent with the other two studies
discussed above.
Per capita oil demand growth is lower than economic growth within the OECD. Vehicle
ownership per capita has stabilized while consumers are starting to purchase alternative fuel and
government policies are mandating higher vehicle fuel efficiency in these wealthier countries.
Studying 96 countries, Gately and Huntington (2002) provided estimates of about 0.55 for the long-
run income elasticity of oil demand in the more mature OECD countries, substantially below an
income response of 1.0 or higher in their study for many non-OECD countries, particularly the fast-
growing Asian economies. Similar results are obtained by Dargay and Gately (2010) when they
update this approach for more recent years. As a result, the rate of increase in oil demand will be
little more than half that of overall economic growth in the developed countries, holding constant
energy prices and technical progress. These OECD income elasticities appear lower than the mean
estimate of 0.93 reported by Graham and Gleister (2004) in their survey of about 100 different road
transportation demand studies (principally gasoline); the survey mainly focused on the more
industrialized countries.
The income responses may be lower in the Gately-Huntington study than in many of the
studies surveyed by Graham and Gleister because the first study allowed the responses to price and
income changes to happen at different rates. They concluded that consumption responds much
more quickly to income than to price, often within the same year as when the economy grows. In
contrast, oil demand adjusts gradually to oil price changes over multiple years, as the economy
replaces its capital stock.
3.2 Oil Demand and Technical Progress
Closely related to the income response is technical progress. This refers to a unidirectional and
irreversible effect on energy consumption that cannot be unwound in the future by a reversal of
price paths or economic growth.
Some progress will accrue exogenously and cannot be attributed to oil price movements or
directly to economic growth. For example, airplanes experienced dramatic increases in fuel
efficiencies well before the 1970 oil price explosions. It is unlikely that airlines will revert back to
earlier airplane technologies if either oil prices should fall or the economy should grow more slowly.
Technical progress may also be closely linked with economic growth. Technical progress at
the aggregate level for an economy may include the shifts in the composition of goods and services
produced in the economy as lighter manufacturing replaces heavier and energy-intensive
manufacturing, and as service activities replace manufacturing. These trends have continued in
many industrialized economies even though aggregate economic growth has subsided considerably
in recent decades. It is very challenging to separate these two effects because economic growth and
technical progress occur simultaneously.
A third source of technical progress relates to past oil price changes if these price increases
are significantly large enough to transform the capital stock. Automobile companies revamped their
entire vehicle fleet after the 1970s to make passenger cars more fuel efficient. When oil prices
declined after 1985, these firms did not scrap their newer automobile designs. In the United States,
vehicle efficiency did decline as these firms sold more trucks where fuel efficiency was not
mandated. But for the most part, the changed vehicle technology was not easily reversed and hence
can be interpreted as price-induced technical change rather than a pure substitution response to oil
price changes. The substitution response implies reversibility regardless of whether oil prices
increase or decrease. Technical progress, on the other hand, implies a unidirectional effect that will
not be easily unwound by new market conditions.
Knittel (2009) studied technological progress in the automobile sector. His findings suggest
that if weight, horsepower and torque were held at their 1980 levels, fuel economy would have
increased by nearly 50 percent from 1980 to 2006 (compared with the 15 percent actual increase).
Knittel’s study also suggests that the technical progress was the fastest during the time that gasoline
prices were the highest.
In the late 1980s and throughout the 1990s, world oil prices declined but oil demand did not
decelerate as quickly as had been expected. Several researchers question the traditional assumption
of perfect reversibility in response to price, as used by Jones (1993), Bentzen and Engsted (1996),
and most of the studies reviewed in previous surveys. Instead, they estimate oil demand as being
asymmetric in its response to prices. Price movements in the 1970s were very large and imposed
“sticker shock” that revamped the economy’s capital stock. Price declines in the mid to late 1980s
did not create the mirror image to these previous shocks. Households adjusted their driving
patterns but did not fully adjust their equipment purchases to this new market environment.
Example studies include, among others, Walker and Wirl (1993), Dargay and Gately (1995, 2010),
Gately and Huntington (2002) and Ryan and Plourde (2002).
Since both price-induced and exogenous technical change can exist, it is not surprising that
different researchers will disagree on these trends. Responding to the 1970 experiences, many
researchers applied an asymmetric-price approach that: (1) separated the 1970 shocks from other
price changes and (2) allowed for different responses between price increases and decreases. Griffin
and Schulman (2005) disagreed and attributed significant time effects (through the inclusion of time
“dummy” or indicator variables) as exogenous technical change. They preferred an approach that
treats the oil demand response from price decreases to be the mirror image to the response from
price increases. Empirical tests that allow for the possibility of both effects appear to support the
conclusion that both price-induced and exogenous technical progress effects are valid (Huntington,
2006, Adeyemi and Hunt, 2007, and Adeyemi et al, 2010).
Another interesting approach applied the structural time series modeling technique to derive
an underlying gasoline demand trend that includes technical progress and other factors that cannot
be easily measured with available data (Hunt and Ninomiya, 2003, and Adeyemi et al, 2010). This
technique often incorporates a significant underlying energy demand trend effect that appears to
have changed sharply during the 1970s. Hence, the pattern of this effect on gasoline consumption is
similar to the price-asymmetric approach, but the technique ascribes no role for the 1970s oil price
shocks in this reversal of the trend effect. An unresolved issue is whether this trend effect captures
too much of the variation in gasoline consumption, because its inclusion can eliminate any long-run
price or income effects in the equation.
One of the challenging issues is the separation of long-run exogenous technology trends
from economic growth (Beenstock and Willcocks, 1981, 1983, Kouris, 1983) and the long-term
impacts of the 1970 oil price shocks. Huntington (2010b) offered a synthesis where he decomposed
the separate effects of oil price movements, exogenous technical progress, income effects and other
time-related factors, all of which move jointly over time. Clearly, more research is required on this
very important issue that may play a critical role in future oil markets.
3.3 Alternative Vehicles and Competitive Fuels
Limited direct evidence exists from past trends on fuel substitution responses to competitive fuels.
Gasoline (and diesel) appears to have no credible competitor for fueling automobiles in many
countries. That experience may be changing as countries like Pakistan are making major
commitments to vehicles fueled by compressed natural gas, and Brazil advances its flexible fuel
vehicles that can use gasoline or ethanol. Longer term, improvements in the batteries for electric
vehicles and plug-in hybrid vehicles may make major in-roads into the gasoline market.
Additionally, companies may decide to build new gas-to-liquid processes to use relatively
inexpensive natural gas to produce diesel fuels. At the moment, economists do not have reliable
estimates on how rapidly gasoline demand would decline for any given decrease in the price of
operating a vehicle based upon one of these competitive energy sources.
Substitution away from oil products and towards natural gas, electricity and biofuels will
depend very much on the relative costs of using the alternative versus oil. It is unlikely that the
relative fuel prices alone will be sufficient to capture this response. Decisions about new vehicle
types will also depend upon vehicle costs, speed, power and other attributes as well as infrastructure
requirements and how the vehicle will be used (local commuting versus longer trips). Additionally,
government mandates and the precise nature of subsidies may be important in encouraging
households to adopt certain energy forms.
3.4 Demand Response to Oil Prices
If future supplies are scarcer, future oil prices will rise and may shave some demand growth. This
issue has attracted probably more attention than any other from energy economists over the last
several decades. Dahl and Sterner (1991) compared the price and income elasticities of demand for
gasoline from more than 100 studies. ; More than 10 years later, Goodwin et al. (2004) updated these
estimates by comparing the results from 69 newer studies. In a companion piece, Graham and
Glaister (2004) reported results from more than 100 studies.
A major conclusion that applies to most studies is that the longer-run response to gasoline
prices over the next twenty years is several times larger than the near-term response during the first
year. The mean estimate of price elasticity from the Graham and Glaister survey for mainly OECD
countries, for example, is -0.25 in the short run and -0.77 in the long run. The variation in estimates
is extremely wide across different studies depending upon time period, data frequency, region and
modeling and statistical approachs. These elasticities are measured at the retail level for a change in
the gasoline price and not the crude oil price.
Considerably less empirical analysis has been applied to other sectors and uses of petroleum.
Pindyck (1979) reported a table of elasticities for the industrial sector in 10 different OECD
countries that are based upon the joint modeling of multiple fuels. The median of his own-price
elasticities for oil is -0.40 with a considerably larger response in North American than in Europe and
The relevant consideration for OPEC’s market power (to be discussed in section 5),
however, is the size of the price elasticity of residual demand (total demand minus non-OPEC
production) at the crude oil price level (Griffin 1992, Hamilton 2009a, Smith 2009). Price elasticities
measured at the crude oil level will be the same as those at the retail level only when refinery margins
and petroleum taxes increase by the same percent as crude oil prices. When crude oil prices change
but refinery margins and petroleum taxes are fixed, crude oil price elasticities will be smaller than
retail gasoline price elasticities. As an example, crude oil prices were 26 to 65 percent of the full
retail price across the seven major OECD countries in July, 2010, with refinery margins and end-use
taxes accounting for the remaining share (International Energy Agency, 2010). If these non-crude
components do not change for this product, these retail elasticities will be equivalent to a short-run
response to crude oil prices ranging between -0.07 and -0.16 and a long-run response ranging
between -0.20 and -0.50. As crude oil prices rise, as they do in the EIA projections, crude oil price
elasticities will rise and OPEC’s market power will decline (Griffin, 1992). If crude oil prices are
twice as high and the same refinery margins and taxes apply, for example, the short-run elasticities
will increase modestly to -0.10 to -0.20 and the long-run elasticities will increase to -0.32 to -0.61.
Once again, the range of crude price elasticities can be very large depending upon region and time
period, as they were for gasoline price elasticities. Krichene (2002) estimates the response of world
demand based upon annual data (1973-1999) to be -0.02 in the short run, while Kilian and Murphy
(2010) estimate the response of US imports based upon monthly data to be greater than -0.25 in the
short run.
The response of consumption to price is the combined effect of many different decisions.
Following the Fisher and Kaysen (1962) approach used for electricity demand, economists often
separate utilization from stock decisions in the demand for energy. Utilization decisions in the
gasoline market influence traffic activity through the number of miles driven by households
(Gillingham 2011). A sustained 10 percent gasoline price increase reduces total traffic mileage by 3
percent after 5 years in the mean results reported in the Goodwin et al (2004) survey. Over a longer
period, households can increase their response by also turning over their capital stock. Stock
decisions include the number of purchased vehicles and the fuel efficiency of the vehicle stock.
Higher gasoline prices can increase the costs of operating a vehicle as well as induce a substitution
away from less fuel-efficient vehicles, both of which will reduce total gasoline usage. For example, a
sustained 10 percent gasoline price increase reduces the total number of owned vehicles by 2.5
percent in the mean long-run results reported in the Goodwin et al (2004) survey.
The long-run stock response is not only larger than the more immediate utilization response.
The two responses may be influenced by very different kinds of price movements (Huntington,
2010a). The price shocks during the 1970s transformed the automobile fleet as well as shifted
driving and traffic patterns. Price movements since the 1970s have been much more modest and do
not appear to have influenced the vehicle fuel efficiency much.5 In reality, there is no one price
elasticity that always applies. Expectations are critical but they are seldom directly observed.
Households that observe a large price change and think that its effect will persist for many years are
more likely to change their vehicle fleet than when price changes are viewed as temporary and
5 Please see the above section 3.2 for a discussion of the price-asymmetric approach for estimating oil demand that
distinguishes the 1970 price shocks from other price movements.
Oil demand’s response to prices may be declining within the United States and perhaps more
widely within the OECD countries. Hughes et al (2008) estimate lower short-run response to the
U.S. gasoline price for the period of 2001-2006 than for the period of 1975-1980. This finding is
consistent with the research on the sticker-shock effect of the 1970 price shocks discussed above.
In a study of the rebound effect,6 Small and Van Dender (2007) estimate lower responses to U.S.
gasoline prices in later years. They attributed it to the declining importance of gasoline use in
household budgets. As incomes rise, households place more weight on the value of their time and
less on gasoline expenditures.
Published data on refined oil product prices before 1978 are limited to only a few large
OECD countries. These countries are generally those with slower oil demand growth rates, which
would create a selection bias against the faster-growing OECD countries as well as the emerging
economies. Although unpublished data on OECD retail oil prices do not appreciably change the
results for the developed countries (Griffin and Schulman 2005), it is expected that they do for the
emerging economies.
Many countries outside of the OECD maintain large fuel subsidies that can impose a very
important wedge between crude and product prices. Removal of these subsidies would raise fuel
prices and reduce future oil demand. The lack of data and estimates for the emerging countries
means that there is a paucity of reliable information on how these changes would influence oil
markets. Van Bentham and Romani (2009) use unpublished data for end-use petroleum product
prices in evaluating energy demand in the non-OECD countries. They estimate a long-run price
response of -0.2 at price levels for the year 2000, increasing to -0.45 for the price levels of $60.
These estimates are somewhat lower than the OECD response estimated from other studies.
Oil supplies are depleting in many fields outside the Arabian Gulf. Development of unconventional
resources will be very important but the cost, availability, and scale of resources like Alberta's oil
sands are unknown. At the same time, oil supply prospects may improve in certain areas. New oil
may be discovered in relatively unexplored regionslike Brazil and West Africa. Reserve
appreciation in known resource basins remains an important source of new additions (Watkins
2002). Technical progress may significantly reduce exploration and development costs.
Governments may reduce oil supply barriers by rolling back royalty and tax rates on oil suppliers
and by easing constraints on leasing and land use, and there is evidence that the latter is already
happening in the United States, for example.
4.1 Resources and Geological Availability
Oil resources are scattered across the globe in formations with very different characteristics. Based
upon its 2000 world oil assessment, the United States Geological Survey (2005) provided a mean
estimate of 1,689 billion barrels of remaining production of conventional oil and natural gas liquids
from known reserves (including future reserve growth from known fields), plus an additional 939
6 Mandates on vehicle fuel efficiency reduce energy use directly but also may indirectly increase energy use by reducing
the cost of operating the vehicle. The second effect is the rebound effect.
billion barrels from undiscovered resources. These geological estimates are based upon likely
discoveries given oil prices and available technologies prevailing in 2005.
This estimate incorporates considerable reserve growth from previously discovered fields.
However, the study was able to assess only 409 of the 937 petroleum provinces in the world. To
extend coverage to other producing regions that were not assessed, Aguilera et al (2009) use a
variable shaped distribution (VSD) model, a size distribution technique that estimates the cumulative
number of provinces as a function of future volumes equal or greater than a minimum size. They
also allowed for reserve growth in these other neglected provinces based upon the growth observed
for the assessed regions. These researchers estimate that if oil production should grow at 2 percent
per year or slightly faster than historical rates, these future conventional oil and NGL resources
would last for 61 years.
Additionally, there are considerably larger volumes of unconventional resourcesheavy oil,
oil sands, and oil shale—that require a specialized extraction technology to produce the oil or
significant processing before the oil can be sold. Aguilera et al (2009) estimate that the total future
volume for both conventional and unconventional oil would last for 132 years if production
increased by 2 percent per year, more than twice as long as for conventional oil.
These estimates of total oil resources are based upon the US Geological Survey assessment
and are more optimistic about future supplies than those provided by Cleveland and Kauffman
(1991). These researchers statistically estimate oil supply equations based upon the geological
approach developed by Hubbert (1962). The Hubbert curve extracts information from discovery
rates, production rates and cumulative production to anticipate when maximum production is
achieved. The rate of petroleum production tends to follow a bell-shaped curve. Early in the time
period, the production rate increases reflecting the discovery rate and the addition of infrastructure.
Later, resource depletion causes production to decline. The biggest success of Hubbert’s approach
is often considered to be his prediction of the decline in US production in the 1966-70 period
(Hubbert, 1956). It began declining in 1970. However, Hubbert’s prediction about peaks in many
other resources has failed to be true. Advocates usually recognize that economics matter but that
geological constraints are paramount. Kauffman and Cleveland (2001) have combined peak oil and
economic constraints with more costly US production. Studying the 48 lower states, they believe if
the government regulations or production costs were different in the late 1960s the peak predicted
by Hubbert would not have been observed (Kaufmann and Cleveland 2001).
Although Hubbert’s idea appears frequently in the literature, it fails to capture the role of
prices in determining supply and demand. As it is described in this paper, there are many other
factors determining the growth in demand and a Hubbert curve excludes these factors in
representing resource depletion. However some modelers have fitted a Hubbert curve that includes
economic indicators, e.g. Baldwin and Prosser (1988), Horn (2004), Dees et al. (2007). Even when
supplemented by economic variables, however, the peaking phenomenon fails to provide a reliable
indicator of resource scarcity (Smith 2012).
4.2 Resource Costs
For economists evaluating market conditions, resource costs rather than peaking determine whether
scarcity prevails. This section reviews recent estimates of direct resource costs, while the next
section extends the discussion to include the additional opportunity cost of using a barrel now rather
than waiting for a future period.
Geologic estimates do not distinguish between resources that are inexpensive to extract from
those that are much more costly to develop and produce. A useful concept, shown in Figure 3, is a
resource availability curve that represents the total known resource base that could be developed at
successively higher cost levels without specifying whether it would be produced in 2015, 2020 or
some other future year. The preferred concept for analyzing oil markets would be an oil supply
curve, which reveals total production for each price level for any given year. The supply curve looks
exactly like Figure 3, except that each curve shows only the volume that can be produced in each
year. Development of such a relationship requires, however, a number of different assumptions
about when resources are developed and under what conditions.
Aguilera et al. (2009) have derived an availability curve for conventional and unconventional
petroleum resources. They estimate 7 TBOE (trillion barrels of oil equivalent) of conventional
resources including oil, natural gas liquids (NGL) and natural gas to be available in the future with
an average production cost ranging roughly from $1 to $20 per BOE (2006 dollars). Conventional
oil and NGLs represent about 3.6 TBOE of the total petroleum resources, with the remainder being
allocated to natural gas. Beyond that there exists 4 TBOE of heavy oil, 5 TBOE of oil sands and 14
TBOE of oil shale with average production costs usually considerably higher than the comparable
costs for conventional oil. We should note that these estimates were prepared several years ago and
may not reflect the more recent and optimistic views of the potential to extract shale gas, which is
itself a significant provider of NGLs.
Cost estimates for these unconventional petroleum resources are very uncertain. Any long-
run cost estimates are for production expanded to scale and after considerable learning from years
of cumulative experience in developing these resources. They should reflect conditions that would
make investment in the infrastructure and all capital costs profitable over many years. If oil prices
are volatile, they need to incorporate a risk adjustment that compensates the investor for his
fluctuating return. In addition, production of unconventional sources can be extremely energy
intensive, which means that costs can and often do rise as oil prices increase. It is also very
emissions intensive, which means that environmental policies may make unconventional sources
very expensive in a carbon-constrained world (Johansson et al, 2009). This production is also very
capital intensive, which means that costs can also increase under conditions of higher capital costs,
as happened in the 2005-2008 period. Finally, producers must also pay taxes and royalties. Plourde
(2009) suggests a combined federal and province income tax of 28.5 percent for Canadian oil sands
that would need to be added to the direct cost of investment. In many major oil-producing
countries, the effective tax rates (royalties, production shares, plus income taxes) exceed 50%
(Johnston, Johnston, and Rogers 2008).
Source: Aguilera et al (2009).
Figure 3 – Global Cumulative Long-run Availability Curve for Conventional Oil and NGL.
Cost estimates for these unconventional petroleum resources are very uncertain. Any long-
run cost estimates are for production expanded to scale and after considerable learning from years
of cumulative experience in developing these resources. They should reflect conditions that would
make investment in the infrastructure and all capital costs profitable over many years. If oil prices
are volatile, they need to incorporate a risk adjustment that compensates the investor for his
fluctuating return. In addition, production of unconventional sources can be extremely energy
intensive, which means that costs can and often do rise as oil prices increase. It is also very
emissions intensive, which means that environmental policies may make unconventional sources
very expensive in a carbon-constrained world (Johansson et al, 2009). This production is also very
capital intensive, which means that costs can also increase under conditions of higher capital costs,
as happened in the 2005-2008 period. Finally, producers must also pay taxes and royalties. Plourde
(2009) suggests a combined federal and province income tax of 28.5 percent for Canadian oil sands
that would need to be added to the direct cost of investment. In many major oil-producing
countries, the effective tax rates (royalties, production shares, plus income taxes) exceed 50%
(Johnston, Johnston, and Rogers 2008).
With frequent energy price and capital cost changes, it is misleading to provide estimates
under current market conditions. Cost estimates from a range of different recent studies (Bartis et al,
2005, Farrell and Brandt, 2006 and Aguilera et al, 2009) are below 2011 prices in excess of $100 per
In a much earlier assessment of resource costs, Adelman and Shahi (1989) estimate the cost
of developing and operating oil supplies across 41 different countries outside the Communist and
OECD regions with publicly available data on drilling. They applied US drilling cost data by field
depth to estimates of average field depth for other countries to develop country-specific cost data.
They adjusted for the higher costs of offshore wells and for other non-drilling expenditures for lease
equipment, improved recovery and overhead expenses. The transparency of their efforts is
particularly noteworthy, because oil production costs are frequently based upon proprietary data and
clouded by their complexity. Smith and Paddock (1984) also estimate variations in development
costs for a broad sample of countries, where the differences reflect variations between onshore and
offshore operations, as well as the harshness of the operating environment and other factors.
It is very difficult to extrapolate future reserve-addition and production trends from a one-
shot, static geologic assessment of the resource base, even if it includes current information about
resource costs. There exists some evidence that the undiscovered resource base grows over time,
despite the fact that more reserves are produced each year to meet market demands (Lynch 2002).
As discussed below, the producer learns more about the geology and what can be extracted
profitably as a known resource basin is developed. This process of reserve growth can be quite
significant in expanding future supplies. In addition, seismic imaging, horizontal drilling, hydraulic
fracturing and other technology advances often make previously expensive resources cost effective,
even in the absence of a higher crude oil price.
4.3 Extraction of a Depletable Resource
The resource availability curve provides no insight about when to extract the resources that are
initially available. Peak oil advocates argue that the resource use will peak once resource limits are
reached. An economic perspective to the same issue is the theory of depletable resources advanced
by Hotelling (1931).
According to this theory, producers know that the resource base is fixed and know the
volume of oil resources that can be eventually recovered. Rather than running out of physical oil
supplies prematurely, market forces work to raise prices and reduce oil demand over time according
to this framework. Suppliers extract resources as long as market prices equal extraction costs plus
the opportunity costs of selling today rather than waiting until the future. In a competitive market,
producers will earn a rent (price minus extraction costs) that rises with the interest rate.
The Hotelling model provides interesting insights about the decisions and expectations of an
individual producer. The approach seems less applicable for explaining historical oil price
movements. Krautkraemer (1998) concluded that there exists very little empirical support for its
main conclusions about oil depletion and its upward pressure on oil prices. Adelman (1990)
similarly criticized the notion that oil and other mineral resources are being depleted. He argues that
the constraint of having a fixed stock of oil to divide between two or more periods (as used in
Hotelling models) is “irrelevant and non-binding” in an economic point of view since the oil
production is determined by future costs and prices. At its most basic level, producers in this
approach extract resources with no technological change, capital requirements, or uncertainty about
the resource base or demand growth. Moreover, resources do not vary in quality and cannot be
replaced by a close substitute.
Although Hotelling framework has received lots of criticism, especially because the forecasts
of the model have not usually matched the observed data, it should be added that this framework
provides an insightful forward looking setting. Similar to many models in economics, the very basic
framework could become enhanced and more realistic by adding details. One advantage of Hotelling
framework is its flexibility for adding such details. Furthermore, the framework can be used to
discuss the effect of future expectations on today’s decisions. An interesting example is the
observation that competitive production may decrease when price increases. One possible argument
for such cases could be the expectation of even higher rates of price increase in the future.
Some researchers have recognized the simplicity of the original model and have tried to relax
several of the more restrictive assumptions. Combining exploration with extraction, Pindyck (1978b)
argued that most resources like petroleum should be viewed as nonrenewable but not as
inexhaustible. Exploration allows the reserve base to grow. With a larger reserve base, well pressure
does not deteriorate and extraction costs may decline. Eventually, however, cumulative extraction
from previous periods raises extraction costs as depletion effects set in and dominate. Pindyck
argues that this phenomenon may explain the U-shape of finding and development costs over the
long run. Extraction costs decline before they begin rising, in contrast to the simplest Hotelling
model where costs are fixed or increasing.
In a later article, Pindyck (1980) emphasized another role for exploration. Drilling reduces
the uncertainty about the resource base as well as builds the reserve base. Better prior knowledge
about the extent and timing of the future resource base may allow the producer to reduce costs.
Slade (1982) develops a Hotelling-style model for natural resource commodities that includes
technical progress. She solves the resulting differential equations to show that the resource price will
follow a U-shape path: prices decline over time before increasing due to depletion. Although the U-
shape path fits well with metal resources prices, but the results do not match the price trends of
petroleum resources.
Intensive expansion of the existing reserve base refers to activity within known fields that
have already been found and developed. This activity will reduce extraction costs because intensive
drilling makes it easier to find additional resources (without exploration) by taking engineering
measures to alleviate the decline in well pressure. Livernois and Uhler (1987), however, underscored
the importance of separating intensive and extensive exploratory drilling. As exploration increases
and the resource base expands, extraction costs may also rise with the addition of new discoveries at
the extensive margin. Newer discoveries are often more expensive because oil producers focused
their initial efforts on the best deposits that are least costly to develop. Their empirical estimates for
Alberta support their argument that finding and developing costs increase as the reserve base is
Oil supply decisions often incorporate the balancing of two different and opposing factors:
resource depletion and technical progress. More cumulative production may deplete the resource
base and increase costs in future years. Alternatively, improvements in oil extraction processes and
our knowledge about the resource base may expand the resource base and reduce future finding and
extraction costs.
4.4 Oil Supply from Competitive Regions
Producers outside of major exporting countries are considered as competitive price takers. Market
prices must cover the costs of producing the last unit, including both the direct expenses and the
firm’s opportunity cost of drilling for oil rather than engaging in another economic activity. If
resource depletion is a factor, the firm will also want to include the opportunity cost of current
extraction relative to future production. At higher prices, firms can justify exploring for and
extracting more costly resources.
The cost of hiring drilling rigs, purchasing key materials, and taxes and royalties are
important components of total extraction costs. If the cost of these factors rises as oil prices rise, the
net profits for expanding oil output may increase substantially less than the increase in the crude oil
price. Likewise, a tax increase may decrease the benefits of the owners but any decrease in the rate of
exploration in one jurisdiction is contingent on the existence of attractive prospects and acceptable
returns for operating companies in other jurisdictions (Kemp and Crichton, 1979). There is some
evidence that the long-run economic incentives do not improve very much in a market with higher
prices, because taxes and royalties also increase (Glomsrød and Osmundsen, 2005). Often these cost
adjustments operate with a lag rather than immediately. Poor data on oil-related taxes and royalties
and their trends prevent a reliable understanding of how economic incentives change when prices
Another important factor will be interest rates. With lower interest rates, producers value
the future more and produce less today. These decisions collectively cause prices to rise in the
current period. Declining interest rates, however, not only influence the extraction rate of the
known resource base but they may also expand the total resources available for all periods by
encouraging more oil investment (Adelman 1982). Declining interest rates will also produce
feedback effects on the economy, encouraging more industrial investment, faster growth and higher
oil demand.
In surveying the literature on non-renewable resources, Cairns (1990) emphasized the
shifting of focus from eventual exhaustion of a homogenous resource base to the transition to lower
quality resources that are more difficult and costly to find and develop. He noted two separate
bodies of research that have not often been combined. Exploration is both a question of depletable
resources but is also a search process that produces important information and experience. This
latter characteristic ties exploration closely to the research and development (R&D) process.
Technical change can be very important but is often difficult to incorporate explicitly in
empirical estimates. Cuddington and Moss (2001) measured technical progress in petroleum and
natural gas exploration and development with estimates of the number of new technology and
process diffusions that come into widespread industry use in each year. Their approach follows a
methodology outlined by the National Petroleum Council and is based upon citations of different
technologies in the Oil and Gas Journal and Petroleum Engineer International. Greater diffusions
significantly reduce drilling costs and tend to offset cost increases induced by resource depletion
(measured as cumulative production in previous periods). Unfortunately, this technology measure is
costly to construct and has therefore not been maintained on a routine basis.
Refinery capacities and their rate of utilization can also influence crude oil as well as refined
product prices. Many existing refineries want low density crude oils with low sulfur content to
produce lighter products like gasoline. Meanwhile, the market is replacing these higher quality crude
types with heavier and sourer crude oils that require different refinery configurations to produce
lighter products. These shifts in crude types and final product demands create upward pressure on
higher quality crude types (Maugeri 2009), most importantly in the short run before the operating
characteristics of existing refineries can be adjusted.
Cohesion among OPEC members is also an important condition shaping oil prices. Some studies
emphasize competitive behavior or non-cooperative responses due to the interdependence among
large producers (oligopoly) or to a strategy based upon some other criteria than wealth
maximization. Other studies have emphasized cooperative behavior among OPEC members, where
oil-exporting nations operate as a unified entity that attempts to influence prices by controlling
production. Empirical tests have tended to show that their observed behavior is more complicated
and often blends different strategies depending upon market conditions. Nearer term, an expansion
in Iraqi petroleum supplies may require other OPEC members to significantly reduce their own
production to prevent oil prices from falling with the increased Iraqi supplies.
5.1 Conceptual Approaches for Understanding OPEC Behavior
Many major oil exporters are sufficiently large to influence and respond to prices. Although some
of the largest exporters could shape future oil prices by themselves, these countries can have a larger
effect if they coordinate their actions. If they withheld supplies, the major exporters would lose
income on any curtailed output but they would gain from increased income on the amounts that are
exported, provided that prices rose sufficiently. The critical response is the net price elasticity of
demand for the producing nation’s exports, which includes their share of the market and the
response of non-OPEC suppliers as well as the response of world demand to price movements.
The relevant elasticities are measured at the crude oil rather than refined product price level, which
can often be substantially different (Griffin, 1992).
This simple intuition provided the catalyst for numerous studies in the 1970s and early 1980s
that presented competing models of OPEC behavior, sometimes with supporting simulations to
display their implications. Although certain producers may have the potential to influence market
prices, the relevant consideration is whether they actually do exercise their market power potential.
Although many researchers believe that oil prices are set above competitive costs, the difference
between having and exercising market power is a critical issue in evaluating oil markets.
Possible explanations in the literature of oil producer behavior include a range of different
hypotheses: competition between many producers, oligopolistic competition between a few large
producers, a single producer operating like a dominant firm (e.g., Saudi Arabia), pure cartel with
fixed market shares for members, and two- or multi-part cartel where groups can bargain for internal
Competition could be perfect, or at least workably so, if no one producer had a large impact
on oil prices, as argued by MacAvoy (1982). There exists considerable uncertainty about drilling and
exploration costs, particularly outside of the United States. Additionally, when oil is being depleted
over time, the full oil production costs will also include the user cost of producing oil today rather
than waiting for a future period, a concept that is very difficult to measure precisely. Under these
conditions, OPEC members would lack the opportunity to pursue price-stabilizing strategies or
other non-economic objectives because they would be competitors with no influence over prices.
Competition could also be oligopolistic where one or several large producers act strategically
to pursue their own economic interests. The type of optimizing behavior will have large effects on
the outcome. Salant (1976) evaluates a situation (a Nash-Cournot equilibrium) where the dominant
producers choose their best price path given the production path chosen by the "competitive fringe"
rather than a path where one group is manipulating price based on the anticipated reactions of the
other (a Stackelberg equilibrium). He finds that the competitive fringe gains more than the
coordinating members when a cartel is formed.
Another approach for representing OPEC behavior is the dominant firm theory where Saudi
Arabia determines the oil price that other members simply accept in their decisions (Alhajji and
Huettner, 2000). Saudi Arabia according to such view may have followed this approach through the
early 1980s but has altered its strategy over the decades. It initially played the swing producer during
1983-85, offering to meet the residual demand that was left unsatisfied after other producers
supplied what they wanted. When it realized that other members were not restricting their output
to increase the group’s income during the early 1980s, it abandoned this approach by increasing its
own production in 1986. In place of the swing producer role, it adopted a “tit-for-tat” strategy in
response to production by other members that exceeded their quota. Saudi Arabia responded by
increasing its production to lower the crude oil price responding to the increased production of
these other producers (Griffin and Nielson, 1994). Saudi Arabia’s announced strategy today is
different, where it tries to respond to the world oil demand with required supply sufficient to
stabilize the price of oil.
A set of models emphasized the cooperative behavior of OPEC. Pindyck (1978a)
emphasizes the importance of analyzing these incentives dynamically rather than as a one-shot static
view. He analyzes the “pure” cartel solution for oil, copper and bauxite. He finds considerable
incentive for cartelization with income gains of 50-100% for oil and bauxite but not for copper. He
attributed these gains to the cartel’s relatively large share of world output, resource depletion and the
slow supply and demand adjustments that leads to extremely high short-run gains.
Other studies revised this approach to allow OPEC members to pursue different strategies.
Hnyilicza and Pindyck (1976) separated the group into spender and saver countries. The first group
had larger cash needs due to rapidly growing population and other factors. They applied a higher
discount rate resulting in more current production than the saver group. Their results suggest that
the relative bargaining power of each block critically impinges on the best price path when the two
groups can change their shares rather than hold them fixed.
It is very important to understand the nature of the demand for the resource. In studying
the bauxite cartel, Pindyck (1977) included a nonlinear demand curve, which was price inelastic at
very low and very high prices but very price elastic when it reached levels where consumers could
easily shift to substitutes for bauxite. This approach might be very valuable for evaluating oil
markets when prices rise to a level that could cause consumers to replace gasoline with electricity,
biofuels and natural gas. Other studies often use an unspecified “backstop” energy source, which
becomes an infinitely available replacement fuel once oil prices reach a given level.
Other OPEC models are not based on cooperative behavior although many explanations
require major exporters to have some market power. The target revenue hypothesis argues that
exporters were interested primarily in reaching some given amount of revenue for meeting internal
investment. These strategies can lead eventually to “backward-bending” supply curves, where
members may curtail production at higher prices, once their revenue goals are met (Teece, 1982).
The property rights thesis (Johany, 1979; Mead, 1979) holds that the transfer of resources from
private international oil companies to governments reduce discount rates for evaluating oil supply
prospects. International oil companies are hypothesized to use higher discount rates because they
fear expropriation of their resources by the host governments. Shifting from higher to lower
discount rates would cause these governments to restrict current production and raise crude oil
prices. Adelman (1986) disagrees, arguing that host governments appear to operate with higher
discount rates than private investors. Finally, Moran (1982) underplayed economic motivations by
emphasizing that exporters were most concerned about maximizing social and political objectives.
They wanted to reduce internal pressures that threatened their own security and expand their
external influence over regional political developments.
5.2 Exporter Behavioral Strategy
The range of different models about exporters’ behavior, the lack of convincing support for any one
of them, and the fact that the major oil exporters were not seriously harmed by making modest
deviations from their best price path led to several popular rules of thumb. The target utilization
model replaces inter-temporal optimization with a simple behavioral simulation rule that allows the
currently observed exporters’ production and capacity to govern OPEC’s decision about price
during the following period (Fischer, Gately and Kyle, 1975, and Gately Kyle, and Fischer, 1977).
OPEC increases price when the demand for its oil is strong and decreases it when conditions are
weak. When oil production increases towards capacity, prices rise as capacity utilization increases.
Prices soften when this capacity utilization decreases. Behavioral rules are particularly attractive
when underlying market parameters and conditions are uncertain. Without perfect knowledge about
these future conditions, countries cannot choose optimal price paths that would necessarily improve
their position in all situations. In other words, the optimal pricing strategy is not robust across
different possible scenarios. Powell (1990) shows that this behavioral simulation rule has failed to
track oil prices well since 1984. He also observed instability problems with a rule based upon
current and past market conditions but that ignores expectations about future conditions.
Later analyses by Gately (2004, 2007) demonstrates that OPEC does pretty well by holding a
relatively constant share of world oil output rather than expand its share dramatically, as implied by
various projections of the U.S. Energy Information Administration’s International Energy Outlook. He
focused specifically on the share of non-OPEC consumption that is attributable to OPEC exports.
One limitation of his approach is that he compared the net present value of oil income without
accounting for the value of the remaining resource at the end of the 30-year horizon of the study.
5.3 Empirical Studies of OPEC Behavior
The past empirical studies have been inconclusive on the behavior and impact of the decisions made
by OPEC. Results have supported many different explanations depending upon the study; few
theories have been rejected. Sometimes, more than one theory can explain the observed results. For
example, cooperative behavior might cause individual members to shift their productions in the
same direction (in parallel) as other OPEC members. However, it might also confirm competition,
because all producers move simultaneously in the same direction once they observe the same price
changes (Smith, 2005).
In the numerous studies of OPEC behavior, the data often reject pure cartel theories such as
a unified organization, a dominant firm, and a revenue-maximizing entity. However, the data often
can support some hybrid or variant of these theoretical frameworks, e.g., partial market-sharing or
partial revenue-targeting approaches. Analysts choose their favorite hybrid, which the data support;
they seldom test all versions. It is likely that the data can support more than one explanation.
Comprehensive testing of all variants on a common data set would be very helpful.
Griffin (1985) initiated formal empirical testing of OPEC market structure at the country
level using quarterly data from 1971:I to 1983:III. His results favored a loose, partial market-sharing
model for OPEC, where member shares change with prices and the production of the group. This
partial market-sharing organization is consistent with Adelman’s (1980) view that OPEC is a
“clumsy” cartel wobbling between competition and market sharing. According to this view,
members find it relatively easy to follow cartel guidelines when prices rise but much more difficult to
follow them when prices start falling. Griffin rejected a target-revenue model and a property-rights
model. His static models related a member’s oil output to price and other members’ oil output.
Non-OPEC oil producers appeared competitive. Similar results apply to the more recent period
through 1988:IV when prices fell (Jones, 1990). These additional tests suggest that the oil price
declines of the 1980s resulted from OPEC members deliberately adjusting output and were not a
breakdown in OPEC discipline.
Dahl and Yucel (1991) tested competing hypotheses for production decisions for both
OPEC and non-OPEC producers. They concluded that loose coordination or duopoly is most
consistent with OPEC behavior. They find no evidence that OPEC producers target revenues, act
as swing producers or act competitively. They also rejected a dynamic optimization thesis that holds
that current production should respond positively to higher interest rates.
Gulen (1996) favored coordination among the members during the output rationing era
when oil prices began to soften during the 1980s. He argues that OPEC started to act as a cartel in
the 1980s in order to prevent prices from falling further. He estimates a long-run relationship
between each member's output and total OPEC production based upon monthly series for all 13
members from January 1965 to February 1993. Monthly data improves the performance of the tests
by increasing the number of observations.
Al-Yousef (1998) tested two economics models for the Saudi behavior in the oil market for
the period 1976-1996 based on decision making structure at OPEC in each period. The first model
considers a swing producer model for the period 1975-1986 and the second is a market sharing
model (similar to Griffin’s market sharing model) covering 1987-1996. Her modeling results
approved the structure assumed for Saudi Arabia in each case.
It can be quite costly to organize and maintain an organization for controlling production
and to enforce its operating procedures and provisions. When costs are significant, the agreement
may be more of a bureaucratic organization where individual members can adjust their production
to compensate actions by other members (Smith 2005). Smith has tested OPEC’s behavior by
searching for a compensating production e.g., an increase in the production of a member in
response to a decrease in the production of another member. The main advantage of this approach
is being independent of estimations for marginal costs, demand and elasticities. Smith argues that
most of the previous empirical tests are either non-conclusive or misreported.
In their tests of OPEC’s behavior, Almoguera and Herrera (2007) find important switches
between cooperative and competitive behavior during the 1974-2004 period. At times, OPEC
appeared to be a Cournot competitor with non-OPEC producers forming a competitive fringe,
much like how Salant (1976) had represented them much earlier. In a Cournot equilibrium, the
dominant producers choose their best price path given the production path chosen by the
"competitive fringe." At other times, they may be adopting a more competitive posture employing
what Griffin-Nielson call a “tit for tat” strategy. It seems that an appropriate way to describe
OPEC’s behavior is to consider two modes for the behavior and recognize the switches that occur
between these modes.
Although they use different data sets and empirical approaches, both Griffin (1985) and
Smith (2005) rejected the dominant firm theory where Saudi production moves consistently against
the output of other members (Alhajji and Huettner 2000).
Smith (2009) has mentioned two price-controlling strategies for OPEC: shutting in existing
production capacity, and restricting the growth of new capacity by limiting the effort to find and
develop new resources. Smith believes that OPEC has mostly failed at the former, but succeeded at
the latter.
Some studies evaluate the effect of OPEC conferences on market outcomes. Wirl and
Kujundzic (2004) argue that the estimated effect of these conferences and their announcements are
at best weak. OPEC decisions may still have an important effect if sufficient information is leaked
prior to the meetings” and that official conferences hardly reveal any new information.
5.4 National Oil Companies
Most of the world’s oil reserves are owned by national entities totally or partially owned by
governments that coordinate oil exploration, development and extraction of the hydrocarbon
resources in their countries, and in some cases outside their borders. National Oil companies
(NOCs) differ in many respects. There are the NOCs of the net oil importers and exporters. They
differ in their evolution, relation to their governments, accountability, efficiency, international
presence, degree of integration, size etc. In this section, the role of national oil companies in oil-
exporting but not oil-importing countries will be considered.
It is not easy to develop broad conclusions about the role and impact of national oil
companies (NOC). Wolf (2009) argues that National Oil Companies (NOCs) in OPEC and outside
OPEC should be discussed separately. He believes that the NOCs of OPEC seem to be more
efficient compared with private companies due to the quality of their resources. On the other hand,
non-OPEC NOCs are less efficient, in terms of labor and capital efficiency. Wolf also discussed the
fundamental differences in goals, policies and data of NOCs and IOCs that often complicate any
meaningful comparisons. Despite this important qualification, some studies have tried to develop
general impressions of this growing phenomenon.
It is often challenging to distinguish between government policy and government ownership
of the petroleum-producing organization and infrastructure. For example, governments might
impose price controls whether the resource is privately or publically owned. Therefore, some
inefficiencies that might be ascribed to NOCs could be attributed to government policies than mere
government ownership of the NOC. Pirog (2007) argues that many of the NOCs found to be
inefficient are based in less-developed countries and are under pressure to maximize the flow of
funds to the national treasuries or provide energy security to the country.
Many NOCs appear to produce less petroleum output than do private, investor-owned
corporations. The reasons are related to the factors already outlined in the discussion of the
OPEC’s behavior. These organizations may restrict current production for several reasons (Hartley
and Medlock 2008):
They withhold more output because they use higher discount rates than competitive firms.
They do not maximize economic profits alone but instead have other political and social
They operate less efficiently incurring higher costs in producing oil.
Unlike private companies, publicly held companies frequently do not disclose sufficient
information about their operations that would allow a better understanding of their activities.
Constrained by this lack of appropriate data, Eller et al. (2010) compared the ability of government
and private companies to generate hydrocarbon revenues with employees, oil reserves and gas
reserves as inputs. They applied both statistical and linear programming approaches to identify each
organization’s relative efficiency. They concluded that generally NOCs are technically inefficient
because they use more employees and reserves per each dollar of revenue generated by the
organization. In situations where NOCs may be required by government policy to sell more
supplies to subsidized domestic markets, it is unclear whether these lower revenues reveal much
about the inefficiency of the NOCs themselves.
In a survey of NOCs in the Middle East, Marcel (2006) observed that most of the NOCs in
those countries provide cheap energy to industrial and residential consumers in their country. Their
citizens expect these huge subsidies to continue; removing them may hamper economic growth and
cause political reactions and an increase in poverty. The subsidies have caused inefficient
consumption behavior, economies more dependent on oil, and even contraband. However,
considering the population growth in these countries, environmental concerns and effects of
globalization, many of the governments may face difficulties continuing these subsidies for many
years. Therefore, it should be expected that there will be a gradual removal of these subsidies in the
following years, resulting in a reduction of domestic consumption and an increase in exports.
This subsidy reduction has already started. Indonesia announced a sharp rollback of
subsidies in September 2005. Iran’s government has passed a new set of regulations to eliminate all
subsidies by 2015. Nigeria’s attempts to decrease subsidies have often met with angry reactions by
the Nigerian people. In 2004 increases in gasoline prices led to calls for a nationwide strike called by
the Nigerian Labor Congress, a scenario that has been repeated recently.
This section summarizes a number of empirical facts about the short-run dynamics of oil price
during the recent oil price boom-bust cycle.
6.1 Higher Volatility beyond the Explanation of Fundamental Factors
There has been a general increase in oil prices since early 2000s, paralleled by a global commodity
market boom. In 2008, oil price rose sharply from $87 per barrel in January to a peak of $145 per
barrel in July before plummeting to $36 in December. As shown in Figure 4, it has since recovered,
trading between $70 and $85 for most of 2010 and between $96 and $123 in 2011 and above $125
by March 2012.
In the meantime, there has been a marked increase in oil price volatility, especially in the first
half of 2008. The annualized daily volatility during 2008 reached over 100%. This is compared to the
volatility of the periods of oil price shocks in 1986 and 1990-1991 as displayed by Figure 5.
Fundamental factors such as demand and supply shocks seemed not able to offer a sufficient
explanation for the observed short-run movements of oil prices in the eyes of a number of oil-
market observers. 7 In this price surge particularly, there have been no oil supply disruptions as in
the historical price spikes. This absence of a supply disruption is also known as the “oil volatility
puzzle” (Baumeister and Peersman, 2010).
6.2 Closer Link with Macroeconomic Variables and the Prices of Other Commodities
During this period of time, oil prices have fluctuated closely with macroeconomic variables, such as
economic growth indicators, exchange rates of US dollar and the US interest rates. Expectations
about future economic conditions can be very important. Drastic oil price changes typically follow
significant macroeconomic news announcements.
7 Some observers believe that fundamentals can explain the oil price shock as long as short-run price elasticities are
sufficiently small. This line of analysis is discussed below in section 7.
Source: Energy Information Administration 2012
Figure 4Daily Brent Crude Oil Spot Price 2000-2012 (Nominal USD per Barrel).
Source: Fattouh (2010a), p10.
Figure 5Annualized Daily Volatility.
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Nominal Price (USD per Barrel)
Oil prices also move more closely with prices of other commodities. Tang and Xiong (2010)
document the empirical fact that correlations between the returns from crude oil and other
commodities have risen steadily from zero in 2004 to around 40-60% during the 2008-2009 period;
this is presented in Figure 6.
6.3 Flat Forward Curves
Another predominant feature of the recent oil price surge is the flat forward curves for crude oil
futures prices. The forward curve is composed of futures prices with different maturity dates. When
this curve is flat, prices for contracts well into the future are not much different from their levels in
more near-term contracts. Prices for long-dated futures contracts reflect, to some extent the market
participants’ expectation about oil price in the future.
Although the spot price or prices of near-term contracts have moved significantly in recent
years, a similar trend applies for the long-term oil prices, too, indicating that beliefs about long-term
prices have also changed. Historically, although the front-end oil futures prices fluctuated
significantly, prices for the long-dated oil contracts were relatively stable, which were believed to
converge to the long-term marginal cost of oil production. In the recent period since 2001, however,
the forward curves have been much flatter, as shown in Figure 7, than for prior periods. While the
price in later maturities often remained in the $20-$22 range for most of the 1990s and early 2000s,
this trend was broken in 2005 when both near-term and far-term prices began moving with each
other, resulting in parallel shifts in the forward curve (Fattouh, 2010a). Litzenberger and Rabinowitz
(1995) study backwardation in oil markets in the early 90s. They document that between February
1984 and April 1992 the nine months futures price was strongly backwardated 77 percent of the
time and weakly backwardated 94 percent of the time.
There are two implications for the flat oil forward curves. First, we have seen unprecedented
shifts in long-dated futures prices. When prices change near term, they also change substantially in
later maturities. Second, front-end and long-dated oil futures prices move much closer with each
Figure 6 – Relationship Between Oil and Other Commodity Returns.
This figure depicts one-year rolling return correlation of oil with soybean, cotton, live cattle, and copper, together with the
95% confidence interval, in Panels A, B, C, and D, respectively. There is a controversy on whether this increase in
correlation is caused by commodities index investment or fundamental drivers.
Source: Tang and Xiong (2010)
Futures curves show the price of futures contracts with different maturities starting from each point of time. Flat
futures curves give you the opportunity to sign a forward contract with a price close to the current spot price. Note that
these curves do not represent the expected futures prices.
Source: Fattouh (2010a), P11.
Figure 7NYMEX WTI Front-Month Contract and Futures Curve.
7.1 Demand and Supply Shocks
Demand and supply shocks, together with the news and uncertainty associated with these shocks,
remain the primary drivers underlying the short-run oil price dynamics. The shocks have been
magnified by the low short run price elasticity of both oil supply and demand.
7.1.1 Supply
In the short-run, oil supply is quite inelastic, that is, the quantity supplied is not responsive to
changes in market price. There are two ways to add additional supply to the market: from either
production if there is spare capacity or from oil inventory. However, from 2005 to 2008, the oil-
producing countries have been running out of their existing spare capacity, due to the under
investment in oil infrastructures during the past decade in some countries or the utilization of
capacity in the demand run up in 2004-2008. Historical OPEC spare capacity is shown with the
Brent Crude oil price in Figure 8. Both Figures 8 and 9 show the historical inverse relationship
between oil spare production capacity and spot oil prices in two different markets. Figure 9 shows
the inverse relationship between OPEC’s spare production capacity and Brent spot oil prices; lower
surplus capacity generally indicates higher oil prices. Since 2004, the spare oil capacity has been
relatively low.
Figure 8 – Brent Oil Spot Price and OPEC Spare Production Capacity (Source: EIA, 2012).
7.1.2 Demand
The short-run demand for oil is relatively price inelastic, meaning the quantity demanded does not
change much relative to price changes. There are four reasons for this. Oil consumption cannot be
changed quickly and easily until the existing vehicle stock and other oil-using equipment turns over.
Second, in the OECD countries, oil consumption is less responsive to price changes because the
share of people’s energy expenditures to their total incomes kept falling until recently. Third, despite
the rising oil prices, oil demand in developing countries was largely driven by steadily growing
income and industrialization process. Particularly noteworthy is oil consumption in China, which
has been growing at a 7% compound annual rate over the last two decades. And fourth, consumers
respond to crude oil price changes slowly in some countries, and the price shocks of crude oil may
be offset by government subsidies or taxes.
OPEC Spare Capacity (MM BPD)
Brent Spot Oil Price ($/Bbl)
Brent SpotPrice
OPEC Spare Capacity
Figure 9 – Oil Spot Price and Spare Capacity (Source: Büyükşahin et al., 2008).
7.1.3 Equilibrium
If both supply and demand are inelastic to prices, it takes large price increases to return markets to
equilibrium if they get out of balance temporarily. Hamilton (2009a) presented a quantitative model
to illustrate the fact that a relatively small demand shock could explain a large proportion of the
price increase. If the robust economic growth shifted the demand curve right by 5.5 million barrels
per day (mb/d) from 2005 to the first half of 2008, and the price elasticity for demand is 0.06, price
of oil should have risen from $55 to $142 (Figure 10).
As presented in Figure 11, the subsequent economic slowdown in 2008-2009 and the de-
layed demand response to these oil price changes can explain the oil price collapse in the second half
of 2008.
A caveat is that the numerical calculations above are extremely sensitive to the assumptions
about the short-run price elasticity of demand. If instead the elasticity were ε = 0.10, the price
would only need to rise to $97 to prevent global quantity demanded from increasing.
7.2 Macroeconomic Variables
7.2.1 News about economic growth
As shown by Figure 12, oil consumption rises directly with income. As incomes increase and
economies expand, more energy will be used for transportation, heating, and cooling. Recently
developed structural models of the global crude oil market, such as the one developed by Kilian
(2009), imply that the surge in the real price of oil between mid-2003 and mid-2008 was driven by
repeated positive shocks to the demand for all industrial commodities, reflecting unexpectedly high
growth mainly in emerging Asia.
Hamilton’s model (2009a) for describing how small demand shocks could result in large price changes in cases of very
small elasticity.
Source: Hamilton (2009a).
Figure 10Price Increases with a Small Demand Shock.
In similar conditions to the previous figure, economic recession could cause a large decrease in prices with a relatively
small change in supply.
Source: Hamilton (2009a).
Figure 11Price Decrease with an Economic Recession.
Figure 12 – Oil Consumption Rises with Income.
Notes: Data are for 2005. GDP is in current U.S. dollars adjusted for purchasing power parity.
Sources: International Monetary Fund; World Bank; Energy Information Administration; Brown et al. (2008)
Thus news about world economic growth, especially in the emerging markets, contributes to
the short-run oil price fluctuation. Hicks and Kilian (2009) confirmed and quantified such effects.
Rather than inferring demand shocks from an econometric model, they utilize a direct measure of
global demand shocks based on revisions of professional real GDP growth forecasts. They show
that recent forecast surprises were associated primarily with unexpected growth in emerging
economies Markets were repeatedly surprised by the strength of this growth. They find news about
global growth predicted much of the surge in the real price of oil from mid-2003 until mid-2008 and
much of its subsequent decline.
7.2.2 Dollar and interest rates
Both exchange rates of the US dollar and the US interest rates have some influence on oil prices.
The impact of exchange rates on oil prices is mixed, as there are several channels through which
the two variables are related. First of all, as a real asset traded in global markets, oil is priced in US
dollars. When the dollar depreciates against other currencies, the crude oil price in the appreciated
foreign currency gets cheaper and thus boosts demand for oil. To clear the market, the dollar price
of oil must rise accordingly to encourage more supply. Initially, depreciation of the US dollar against
a basket of other foreign currencies would be followed by an oil price increase to offset the decline
in real oil prices in other countries, when all other economic factors are held constant. This suggests
a negative correlation of dollar exchange rate and oil price. The strength of the effect is limited,
however, by the relatively small price elasticity of demand for crude oil.
The second channel for dollar-oil price correlation is the production decisions of key oil
exporters. When the dollar depreciates, oil exporters suffer a decline in the international purchasing
power of their revenue. For this reason, oil exporters, like OPEC, have strong incentive to exert
their market power to lift oil price.
The effect could be in the reverse direction, where the oil price influences the value of the
dollar. In 2008, oil imports accounted for about 49% of the US trade deficit, up from 18% in 2002.
Higher oil prices could lead to larger trade deficits, which might drive the dollar value even weaker.
This effect could be offset by shifts in international asset trades.
Furthermore, both the depreciation of the dollar and the oil price may shift in response to
other economic conditions. Other common factors may also cause correlation between oil prices
and exchange rates. For example, robust economic growth in the US may appreciate the value of
dollar and lift oil prices at the same time. Another example is unexpected interest rate cut by the
Federal Reserve Bank, might lead to a weaker dollar and a higher oil price.
Empirical results suggest the correlation between oil prices and the dollar exchange rates is
not constant but time-varying. According to Medlock and Jaffe (2009), from January 1986 to January
2001, the correlation is -0.08, while from January 2001, the correlation is -0.83. The correlations for
each period are presented in Figures 13 and 14, respectively. This strong negative correlation
suggests that world economic conditions created strong global economic growth with a weaker
dollar without their necessarily being a direct causal link between the two variables.
Data Source: EIA 2012
Figure 13 – Dollar Exchange Rate and WTI Crude Oil Price Through 2000.
Data Source: EIA 2012
Figure 14Dollar Exchange Rate and WTI Crude Oil Price After 2000.
Figure 15 suggests that a weak dollar explains only part of the story. From 2002 to 2008,
U.S. dollar depreciated by nearly 25 percent, while the dollar price of oil rose by more 600 percent.
Clearly, oil prices measured in different major currencies have also risen sharply.
The US interest rate is another important macroeconomic variable for oil prices. One
mechanism by which declines in interest rates could drive up the current period’s oil price is through
a reduction of costs to store oil inventories that could encourage producers to sell less oil today.
Both interest rates and the oil price may shift in response to other economic conditions.
The relationship between interest rates and oil prices is also time-varying, depending on the
interaction of a number of macroeconomic variables. If the decline in interest rates is in reaction to
an economic recession, we should also see the falling oil prices in response to the weak demand. In
this case, we observe a positive correlation. A lower interest rate may be associated with the
expectation of higher inflation rates, especially during the business expansion, and this expectation
typically pushes up oil prices, suggesting a negative correlation.
Interest rates are important for those trading on the financial markets. Gracia (2006) cited
low interest rates as one of the classical indicators of a potential speculative bubble as that lowers the
funding cost of financial traders. Also, lower interest rates give financial investors higher incentives
to shift money from low-return treasury bills to high-return commodity futures.
7.3 Financialization and Speculation
Since 2004, there have been growing “open interest” and trading volumes in crude oil and other
energy futures. This development is part of a broader financialization process where the vastly
expanded role of financial motives, markets and institutions have dominated commodity trading.
Tang and Xiong (2010) document the increased correlation among different commodities and link
this finding to an increase in commodity index investments. Buyuksahin and Robe (2010) study the
correlations between the returns on commodity and on equity indices. After controlling for
macroeconomic fundamentals, they show that hedge fund activities can help explain commodity-
equity co-movements observed in recent years. They also point out, however, that there is no clear
indication of what a “normal” correlation between commodity and equity prices might be; i.e., no
indication of whether a high correlation is “good” or “bad.”
Source: Interagency Task Force on Commodity Markets (2008)
Figure 15Exchange Rate and Oil Prices in Several Currencies.
As shown in Figure 16, the total open interest in WTI crude oil futures, measured in the
number of oil futures long or short contracts that have not been settled (open interest), has more
than tripled since 2000, with the peak in June 2008.
Oil futures and options are mainly traded through organized and regulated exchanges, such
as the New York Mercantile Exchange (NYMEX). Traditionally, in the organized exchanges of
NYMEX and IPEthe activities of basically two categories of market participants in the oil market
have been reported: (1) commercial traders; (2) non-commercial traders. An even larger volume of
forward contracts are arranged on the overthecounter market (OTC). Financial economic research
focuses exclusively on futures and options because no OTC data are available for these unreported
and unregulated transactions. Commercial traders include both oil producers and oil consumers.
They trade oil futures to hedge the risk of oil price movements that are unfavorable for their
ongoing business activities. Oil producers tend to hold net short positions in oil futures while
consumers are more likely to keep net long positions. For commercial traders, in order to hedge oil
price risk, they need to find the potential counterparties to sign futures contracts in any market
circumstances. Non-commercial traders serve this role and provide liquidity to the futures market.
Typically, they trade frequently and actively with the aim to gain from the short-term price changes.
Non-commercial traders include hedge funds, managed futures funds, commodity trading advisors
(CTA) and other financial institutions. They seek to make profit on paper positions from oil price
Figure 16 – Total Open Interest (Thousands of Contracts)
Open Interest refers to the total number of futures contracts that are not closed or delivered on a particular day. For
each buyer of a futures contract there must be a seller. From the time the buyer or seller opens the contract until the time
it is closed, that contract is considered 'open'. A large open interest indicates more activity and liquidity for the contract.
A common misconception is that open interest is the same thing as the volume of futures trades. Buying a futures
contract from someone who already owns one increases the volume of trades but not the open interest.
Source: Fattouh (2010a), P30, Figure 12.
7.3.1 Financialization and index investors
In recent years, a new group of passive index investors, who are mainly institutional investors that
include pension funds, mutual funds, and endowment funds, have emerged as important players in
crude oil and commodity futures markets, as shown in Figure 17.
The investment objective of a commodity index trader is to track an index of commodities
usually the Standard and Poor’s S&P GSCI (formerly the Goldman Sachs Commodity Index)8 or the
Dow Jones UBS index9or a sub-index of one of these over time by acquiring long positions via
bilateral or over-the-counter (OTC)swap contracts, index funds, or exchange-traded futures. The
larger commodity index traders typically gain commodity exposure through swap dealers, who
exchange the maturity and other traits of one security for another. The investor usually takes the
long position of the index while the swap dealer is short. The swap dealer, which is often affiliated
with a bank or other large financial institution, has emerged to serve as a bridge between the OTC
8 This tradable index is readily available to market participants of the Chicago Mercantile Exchange. Investors in the
commodity markets use it as a benchmark for measuring commodity performance over time.
9 Dow Jones UBS index, based upon UBS investment bank, includes commodities traded mostly on U.S. exchanges in
addition to aluminum, nickel and zinc that are traded on the London Metal Exchange (LME).
swap market and the futures markets. They offset their aggregate risk exposures through buying or
selling commodity futures and options. Figure 18 is a schematic showing the relation between the
institutional and individual demand for commodity index portfolios and the supply of commodity
index portfolio replication contracts as provided by the commodity futures market.
The S&P GSCI (formerly the Goldman Sachs Commodity Index) is a tradable index that serves as a benchmark
for investment in the commodity markets and as a measure of commodity performance over time. The index contains a
much higher exposure to energy than other commodity price indices such as the Dow JonesAIG Commodity Index.
The DJ-AIG commodity index is composed of futures contracts on physical commodities. In DJ-AIG no one
commodity can comprise less than 2% or more than 15% of the index and no sector can represent more than 33% of
the index.
Source: Masters (2008)
Figure 17 – Growth of Passive Index Investors.
Source: Master and White (2008), p10.
Figure 18 – Commodity Index Portfolios and the Commodity Futures Market.
Institutional investors’ interests in commodity index were mainly motivated by the fact that,
over a long period, investment on crude oil and other commodity futures delivered equity-like return
and risk, while their returns are uncorrelated or negatively correlated to conventional asset classes.
Thus commodity futures could potentially be a good addition to the overall asset portfolio to lower
risk and enhance return; see the work of Gorton and Rouwenhorst (2004). Moreover, because
commodity returns are often positively correlated with inflation, it is possible for investors to invest
in commodities as a means to hedge against rising inflation.10
Figures 19 and 20 show the increases in both commercial and non-commercial activity
from 2003 to 2008. Among the non-commercial participants, both hedge funds and floor brokers
and traders exhibit robust growth in open interest. Among commercial traders, much of the growth
in open interest comes from greater activity by two categories commodity swap dealers and
commercial dealers. While commercial dealers utilize futures trading to manage price risk for the
purchase and sale of physical commodities, commodity swap dealers use futures markets to manage
price risk stemming from their OTC swap business and also to handle the majority of commodity
index trades in the futures markets.
In 2008, the U.S. Commodity Futures Trading Commission (CFTC) published a report on
commodity swap dealers and index traders swap dealers to give a snapshot of market participants in
the WTI crude oil futures and other commodities. The estimated aggregate net amount of all
commodity index trading (combined OTC and on-exchange activity) on June 30, 2008 was $200
billion, of which $161 billion was tied to commodities traded on U.S. markets regulated by the
CFTC (Figure 21). The net notional index value for futures and options open contracts on June
30th, 2008 for NYMEX crude oil was the $51 billion, which accounts for 13 percent of the total
notional value of $405 billion for NYMEX futures. The report shows that for commodities such as
wheat, over 90% of the swap dealer long positions represent index traders; whereas, in crude oil it
may be as low as 41%. Figure 22 shows that the open net position for oil financial investors did
not jump when oil prices spiked in 2008.
10 Institutional investors could play a stabilizing role. When oil prices rise, the share of commodities in their overall
portfolio increases which induces institutional investors to sell commodities in order to rebalance the portfolio pushing
prices down.
Non-commercial participants are generally those with no interest in the physical market.
Source: Fattouh (2010a)
Figure 19WTI Average Open Interest by Non-Commercial Participants, 2003-2008.
Commercial participants are generally those with interest in the physical market.
Source: Fattouh (2010a)
Figure 20WTI Average Open Interest by Commercial Participants, 2003-2008.
Source: U.S. Commodity Futures Trading Commission (CFTC 2008)
Figure 21 – Index Fund Values and Shares of Open Interest, June 2008.
Non-commercial market participants are those with no interest in the physical market. Therefore their position in the
market may be considered as speculative behavior. This figure shows that there is no common trend between these
actions and oil prices. Furthermore, the changes in non-commercial positions are much more than the changes in oil
Source: Fattouh (2010a), p35.
Figure 22Net Position as a Percentage of Open Interest (Non-Commercials) and Oil Price.
7.3.2 Is there a speculative bubble?
Have the financialization and the increasing presence of index funds caused a speculative bubble in
oil prices? Brunnermeier (2009) defines a bubble as “asset prices that exceed an asset’s fundamental
value because current owners believe that they can resell the asset at an even higher price in the
future.” A speculative bubble is characterized by the following elements: (1) Prices are higher than
the fundamental value; (2) A group of investors buy the asset based on the belief or sentiment that
they can sell it to others later with a higher price; (3) Such beliefs or sentiments cannot be supported
by fundamental factors.
Masters (2008), along with the Staff Report (United States Senate 2006) by the Permanent
Subcommittee on Investigations of the Committee on Homeland Security and Governmental
Affairs, ascribed the rapid increase in overall commodity prices from 2006-2008 on institutional
investors’ embrace of commodities as an investable asset class. He made an analogy to the activity of
index funds as to the infamous Hunt brothers’ cornering of the silver market. Masters and White
(2009) recommended specific regulatory steps to address the alleged problems created by index fund
investment in commodity futures markets, including the re-establishment of speculative position
limits for all speculators in all commodity futures markets and the elimination or severe restriction of
index speculation.
Although the speculation argument is appealing, there still exist a number of inconsistencies
with the story. Irwin and Sanders (2010) have summarized arguments against the bubble story in
crude oil as well as the general commodities markets.
The first conceptual error of the bubble argument is equating money inflows to commodity
futures, i.e. paper demand with physical demand. The examples include Masters (2008) and US
Senate (2006). They made the wrong analogue between the rising prices with some historical
commodity futures bubbles (for example, Hunt Brothers in silver market in 1970s), in which there
were manipulation in spot markets.
Second, if prices are to be distorted by noise traders, their behaviors must be unpredictable;
because otherwise, rational traders will exploit the riskless arbitrage to make money and at the same
time bring the price to the equilibrium level. Actually, the rolling schedules of most commodities
indexes are fixed and quite predictable.
Third, price movements in futures markets with substantial index fund investment have not
been uniformly upward; see Irwin et al. (2009). Also, Headey and Fan (2008) show that prices of the
non-financial commodities that are not part of any commodity index show similar dynamics as
commodities in the index (rubber, onions, iron, apples and edible beans). If index investing is the
main driver behind the price surge, we should not expect to see similar price increases in such non-
indexed commodities (Figure 23).
Fourth, Till (2009) uses a Working’s T index, a traditional metric for evaluating speculative
position taking, and finds that this position-taking does not appear to be excessive over the past
three years when compared to the scale of commercial hedging during 2006-2008 (Figure 24). This
study provides further evidence that excessive speculative behavior is not evident from the
observable data.
Sanders et al. (2008) define Working’s T index as follows:
T = 1 + SS / (HL + HS) if (HS >= HL), or
T = 1 + SL / (HL + HS) if (HL > HS)
where open interest held by speculators (noncommercials) and hedgers (commercials) is denoted as
SS = Speculation, Short
HL = Hedging, Long
SL = Speculation, Long
HS = Hedging, Short
This figure displays an equally-weighted index of spot prices for seven non-exchange traded agricultural and industrial
commodities. The index covers rice, coal, and five industrial metals: manganese, rhodium, cadmium, cobalt and
tungsten. To abstract from absolute price-
level differences across these commodities, the index is computed by
compounding an equally-weighted average of the individual commodity returns, and deflating the resulting series.
Non-Exchange Traded Commodity Real Price Index (1990=100)
The spot price data are from Bloomberg; price indices from the US Bureau of Labor Statistics.
Source: Buyuksahin et al 2008
Figure 23Commodity Fundamentals.
Source: Till (2009), p10.
Figure 24 – Working T Index for June 2006 through October 2009.
7.3.3 Krugman’s argument regarding inventory stockpiling
As Hamilton (2009b) pointed out, the difficulty regarding the speculation story is how to reconcile
an explanation of oil-price determination as a financial bubble with what is happening to the
physical quantities of the product that are produced and consumed. In order to have an impact on
the spot oil prices in the cash market, speculators or index traders would have to take delivery and
buy quantities in the cash market and hold these inventories off the market. Alternatively, they
would have to convince commercial entities of an impending shortage through their trading activity
and thereby induce the commercial entities to withdraw oil from the commercial supply chain. In
fact, index investment rarely got involved in the cash market activity except for some anecdotal
evidence. Krugman (2008) contends that, for storable commodities, if speculation is to raise price
above the fundamental price, it follows that supply exceeds consumption. The excess supply will
automatically lead to the increased crude oil inventory. Consider a very simple model to illustrate the
point. Using the notation of Einloth (2009), let Q(P) denote the oil production function during a
certain period of time, () the demand function, and the aggregate oil inventory. As a
consumption commodity, the fundamental price of crude oil, , is defined as the level at which
quantity supplied equals quantity consumed. For simplicity, here we neglect the effect of seasonal
inventory adjustment. Thus, ()=(). If oil price is raised by speculation to higher than
the fundamental level, there must be excess oil supply, as ()<() . The supply surplus
must be accumulating as inventory. For any price level that clears the spot market, it must have:
Figure 25 illustrates the point.
For speculation to have a lasting effect on oil price, oil inventory must continue rising. This
suggests a simple test for the effect of speculation on oil prices: verify that there is a continuing
increase in oil inventory.
Figure 25 – Excess Supply Leads to Increased Crude Oil Inventory.
However, oil stocks were declining in 2006-2008 instead of growing. As illustrated by
Figure 26, the U.S. oil inventories were unusually low during oil price run-up between May and July
of 2008. Similarly, there is no evidence of a substantial increase in OECD petroleum inventories
since 2000, as shown in Figure 27. Now, two questions arise. First, if there was a bubble in which
spot oil prices were higher than the fundamental prices, where was the missing oil inventory? Second,
how much increased inventory was needed to explain the oil price run-up from $35.
Hamilton (2009a) argues that the oil price surge in July 2008 was influenced in part by the
flow of investment dollars into commodity futures. But he emphasized that the two key ingredients
needed to make such a bubble story coherent include a low price elasticity of demand and the failure
of physical production to increase. With very low short-run demand elasticity, it would take many
months before any excess supply could be observed in the public inventory data. A small amount of
inventory buildup may be enough to sustain a bubble price. Hamilton (2009a) further pointed out
that “this condition-- extremely low price-elasticity of demand-- is precisely the same condition
under which it is not too hard to make sense of observed oil price changes based purely on
fundamentals alone.” Pierru and Babusiaux (2010) developed a dynamic model in which demand
elasticity is an increasing function of the time horizon. Results from calibrating this model suggest
that speculation may temporarily push crude oil prices above the level justified by physical-market
fundamentals, without necessarily resulting in a significant increase in oil inventories.
Source: Energy Information Administration
Figure 26 – Crude Oil Inventory in the United States.
Kilian and Murphy (2010) have constructed a structural vector autoregressive (SVAR) model
that includes oil production, inventory, price and a proxy for real economic activity. A well-known
problem of the conventional econometric models about the oil system is the endogeneity issue, that
is, two or more variables are determined jointly by a set of other factors. This problem means that
one or more variables on the right side of the regression equation are correlated with the error term
and standard single-equation estimation techniques may not apply. Using a recently developed
method called “sign restrictions”, Killian and Murphy were able to distinguish supply shocks,
demand shocks, and speculative shocks and thus identify the structural parameters, to some extent.
The power of this approach depends very much on the validity of the restrictions imposed by the
model. This identification is not conventional point identification but the set identification, that is, a
small of number of admissible models are identified. Their estimate for the short-run demand
elasticity is -0.26, much larger than earlier studies using models without accounting for price
endogeneity. This result rules out the case of zero short-run demand elasticity conjectured by
Hamilton (2009a). In addition, they find that the 2003-08 price surge was mainly caused by oil
demand driven by the business cycle. Speculative shocks, which did play important roles in 1979,
1986, and 1990 oil price shock episodes, contributed little to the price run-up after 2003 or even the
price surge in 2007-08.
NOTES: The OECD inventory data have been normalized such that the inventory value in logs is zero in
2000.1. Source: Kilian (2010)
Figure 27OECD Petroleum Inventories and the Real Oil Price.
The second problem of checking oil inventory is the difficulty in tracking oil consumption,
production and stocks accurately. The practice is more complicated than the simple model as a
proportion of inventory changes may also be induced by fundamental reasons. For example,
producers might add to inventory in anticipation of high demand during the summer driving season
or winter heating season. In addition, oil could be stockpiled somewhere in the world, but not
necessarily by U.S. or OECD oil refiners. It is impossible to check inventories in Asian countries or
in tankers at sea. For these reasons, it may not be easy to identify the inventory buildups due to
The third possibility to explain the puzzle of the lack of accumulating above-ground
inventories is underground stockpiling. Davidson (2008) argues that the absence of higher
inventories does not necessarily indicate the absence of excess speculation in the market. If oil prices
are expected to rise in the future more rapidly than current interest rates, then commercial producers
like OPEC and multi-national oil companies, can enhance total profits by leaving more oil
underground today for future production. Using earlier notation, speculation reduces current oil
supply Q(P) as opposed to increasing inventories. In this sense, the concept of fundamental prices
that equate current supply and demand is not sufficient to capture expectations in a dynamic setting.
Smith (2009) finds contrary evidence for U.S. producers during 20072008 when they were trying to
increase production in the face of rising prices. This issue deserves more attention and needs to be
investigated for a wider group of producers.
Expectation of a sustained high oil price in the future rationalizes what Kilian (2010) calls
“passive speculation by producers”. Long-dated oil futures price reflects such expectation to some
extent, although not perfectly. Parsons (2009) cited the flat oil forward curves in recent years as
evidence in favor of this argument. For example, when spot oil price reached over 140 dollars per
barrel in July 2008, futures prices for oil contracts that expired from two to five years were
uniformly close to 140 dollars per barrel. As Fattouh (2010b) pointed out, this argument highlights
the role of expectation about oil demand and supply fundamentals in the future. When a producer
expects he can sell oil in the next few years at a higher or similar price, it is optimal for him to leave
oil under the ground now and sell it later. It is hasty to ascribe such expectation to speculation as
such expectation could largely be driven by the fear of tight supply coupled with robust demand in
the future.
7.3.4 Evidence from statistical tests
Statistical methods have been applied to test if there was a speculative bubble in the oil market
caused by financial investors. So far, they have failed to find significant statistical evidence for a
conventional speculative bubble or market manipulation, except for some weak evidence for a
potential bubble during the first half of 2008. Granger causality test
It is argued by some researchers that if prices were driven up by index investment or other
speculators, those net positions should lead to price changes. A number of empirical papers use
Granger causality tests to evaluate the impact of speculators’ position on oil price and volatilities.
Granger causality tests using a variety of data and variables failed to provide such significant
Granger (1969) suggests the bivariate Granger test for examining the lead-lag or “causal”
relationship between two time series. Granger causality test is a standard linear technique for
determining whether one time series is useful in forecasting another. A time series X is said to
Granger-cause Y if it can be shown, usually through a series of tests on lagged values of X (and with
lagged values of Y also known), that those X values provide statistically significant information
about future values of Y.
Typically, the following models are estimated to evaluate the role of speculation on oil
futures returns (or volatilities) and vice versa.
where represents oil futures returns or volatilities and  represents a change in financial
position. A variety of measures for speculative positions or index positions have been used. The
numbers of lags are selected according to some information criteria that evaluates whether the
number of additional parameters (lagged values) sufficiently improve the equation’s goodness of fit.
If the null hypothesis of no causality from position to returns (H:= 0) is rejected at the 5% level,
the test concludes that positions Granger cause returns. The test is whether the variance of residuals
from the respective unrestricted model (which includes lagged financial position terms) is
significantly less than the variance of residuals from the corresponding restricted model (which
excludes lagged financial position terms).
The most widely cited study using Granger causality test is the Interim Report on Crude Oil
(Interagency Task Force on Commodity Markets, 2008). The study used non-public data that
examines the dynamic relation between daily price changes of nearby NYMEX WTI crude oil
futures and positions of various categories of traders from January 2000 to June 2008. Commercial
traders are divided into manufacturers, commercial dealers, producers, other commercial traders and
swap dealers. Non-commercial traders are divided into hedge funds and floor brokers & traders.
Position changes are defined by net daily positions changes in the nearby futures plus delta-adjusted
options contract.
The study found no systematic evidence that changes in net positions of any subsequent
category of traders lead to oil price changes, as presented in Table 1. Instead, commercial traders,
especially manufacturers, commercial dealers and producers, adjust their positions in response to
price changes over the sample period. Specifically, they increase their positions when prices rise and
lower their positions when prices fall. Although hedge funds are found to follow price changes,
non-commercial traders as a whole do not respond to price changes.
Buyuksahin and Harris (2009) updated and enhanced the results of the previous study in
several aspects. First, the sample period was extended to March 2009, covering the price collapse in
late 2008 and price recovery in early 2009. Samples are divided into two subsamples: 2000-2004 and
2004-2009. Their sample frequency ranges from daily to five days. The results are in line with those
in the previous study.
These two studies use non-public data on daily frequency. Irwin and Sanders (2010) use
weekly announced Disaggregated Commitments of Traders (DCOT) data published by the U.S.
Commodity Futures Trading Commission (CFTC) for crude oil and natural gas. Based upon this
other data, they have reached a similar conclusion.
A number of studies applied Granger causality test to other commodities futures, including
Stoll and Whaley (2010), Irwin and Sanders (2010), Sanders and Irwin (2010) and Sanders and Irwin
(2011). To improve the power of the traditional Granger causality test, Sanders and Irwin (2011)
adopted an alternative econometric method, seemingly unrelated regressions (SUR), which considers
the cross-sectional correlation of the error terms. Most of these studies give similar results: some
sub-categories of traders follow price changes but positions of any sub-category of traders do not
lead price changes. Different from most Granger causality tests that focus on time series comparison,
Sanders and Irwin (2010) implemented a cross-sectional comparison that also did not support the
significant impact of index funds on futures price for twelve agricultural futures.
Despite the consistent findings, there are several caveats regarding Granger causality test and
the existing results. First, Granger causality does not imply true causality. If both X and Y are driven
by a common third process with different lags, their measure of Granger causality could still be
statistically significant. To develop the true causality relationship, an economic model is often
needed. Second, poor data accuracy may lower the power of the test, i.e., the probability of rejecting
the null hypothesis when the causality does exist.
Table 1Granger Causality Tests for WTI Crude Oil Futures.
Source: Interagency Task Force on Commodity Markets (2008)
Another major problem of Granger causality test is the difficulty of choosing the right
number of lagged periods. Many Granger causality tests are based upon returns and traders’
positions over a very short horizon, for example, one day or one week. Although Granger causality
test may well capture the short-run influence of investor activity on oil prices, it may not be able to
detect the case in which the flows of index investors or other trader categories affect asset prices
gradually over longer horizon than just a few days.
And finally, whatever contribution speculators may be making, arbitrageurs will always seek
to eliminate predictable changes. Hence we might generally expect to find that nothing Granger-
causes returns even if non-fundamental factors are making a contribution. For example, standard
models of a stock market bubble still maintain that stock price changes are unpredictable.
Convenience yield and speculation
Section 7.3.3 above explained the difficulty of detecting an oil price bubble using inventory data. In
order to avoid this problem, a number of studies use an alternative concept convenience yieldto
address this issue. The convenience yield, defined as the marginal value of holding one unit of
physical commodity inventory, has long been used in research on storable commodities. These
benefits include the ability to profit from temporary shortages, and the ability to keep a business
operation running. Such benefits are not obtained from holding the futures contract. The theory of
storage by Brennan (1958) and Telser (1958) documents a negative relationship between the
convenience yield and inventory. Intuitively, when inventories are high, this suggests an expected
relatively low scarcity of the commodity today versus sometime in the future.
It is easy to see that the convenience yield accrued by inventory holders is analogous to
dividends on stocks. Pindyck (1993) proposed a present value model to link the commodity spot
price to the flows of convenience yields. The present value model is written as,
S= E[e()
where S is the current spot oil price, and is the flow of net convenience yield at time t. Like the
present value model for stocks with dividends, this equation is treated as a benchmark for rational
asset pricing problem. Similarly, the current spot price may deviate from the level implied by
present value model if the market is not rational or affected by speculation. He also shows that if
the present value model holds, there should be a cointegration relationship between the spot price
and the convenience yield, that is,
S= a + b+
where is a stationary process. Although the convenience yield is not observed directly, it can be
inferred approximately from the no-arbitrage relationship between spot price and futures price by
=(1 + r)SF(t, T)
where F(t, T) is the futures price at time t with maturity T.
Liu and Tang (2010) test the cointegration relationship between the spot price and the
implied convenience yield for four commodities. They find that, for crude oil before 2004, the spot
price and the convenience yield are cointegrated with each other, while after 2004 the cointegration
relationship does not exist. Thus, there exists a structural break around 2004, which is consistent
with the findings of Tang and Xiong (2010). Oil prices after 2004 seem beyond the levels that can be
justified by market fundamentals represented by the convenience yield. To develop a direct link
between this gap with the financialization process in commodities markets, the authors constructed
a theoretical model in which both commercial traders and speculators are explicitly introduced. The
first type of agents own physical inventories and write commodity futures. The second type of
investors only invests in paper-based futures but they also hold a position in the stock market. They
have shown the current spot price can be written as the linear combination of current convenience
yield and risk premium in the stock market. Comparative statics analysis shows that as more and
more financial investors enter the commodity futures market, the spot price gets higher and it moves
more closely with the risk premium of the stock market. Even if the spot price is above the
fundamental prices, the financial investors are still willing to buy the futures since they can achieve
the benefits of diversifying their portfolio risks. Their findings are consistent with Etula (2009) who
links the returns of commodity futures to the risk appetite of broker dealers.
Einloth (2009) recovers the convenience yield from futures prices with different maturity
dates and studies the co-movement of oil prices and convenience yields. The findings of the study
are shown in Figure 28. He argues that the changes in the convenience yields combined with
changes in oil prices help distinguish the speculative demand and precautionary demand. Under the
fears of supply disruptions, the marginal convenience yields rise given the quantity of inventory,
providing more incentives to hold more inventories. The build-up process continues until the
implied marginal convenience yields are close to the marginal storage costs. In contrast, a pure
speculative demand shock also induces a rise in inventory, followed by a decrease in the marginal
convenience yields. The author contends that from early 2007 to March 2008, both spot prices and
convenience yields increase, indicating a positive demand shock and inconsistent with speculation
having played a role. In contrast, from March to July in 2008, spot prices rise further but
convenience yields started to fall, suggesting that the rise in demand during this brief period was at
least partly that of speculators building inventories. From August 2008 through February 2009, both
oil prices and convenience yields fell rapidly due to a negative demand shock, which is consistent
with the fundamental story.
The convenience yield is defined as the marginal value of holding one unit of physical commodity inventory. This
figure presents implied convenience yield based on futures prices of different maturities. There are some times that
prices and convenience yield move in the same direction; this could indicate a demand shock.
Source: Einloth (2009), P17.
Figure 28WTI Price and Convenience Yield. Explosive prices in a bubble
Using the econometric test developed by Phillips and Yu (2009), Gilbert (2010) examined whether
crude oil prices exhibit the time series pattern of a series in an explosive bubble. The underlying idea
for this test is simple. If trend-following behavior is important, an upward movement in prices will
tend to be extrapolated. The view that there is a strong (positive or negative) trend in an asset price
will itself generate the momentum which validates this belief. The author applied this test to the
futures prices of WTI crude oil, aluminum, copper, nickel, corn, soybean and wheat from January
2000 to June 2009. Using monthly average prices, he finds statistical evidence for an explosive
bubble for copper over the eight-month period February to October 2006 and bubbles during two
isolated months for nickel, but not for WTI crude oil. Using daily oil prices instead, he finds only
weak evidence for an explosive bubble that last a few days in July 2008.
In summary, most of the studies surveyed found no significant statistical evidence for a
speculative bubble in the 2004-2008 oil price boom except for some weak evidence during the first
half of 2008. In particular, the direct link between the oil price movements and commodity index
funds have not been well developed in existing empirical studies.
7.4 Financialization as a Double-Edged Sword
7.4.1 Financial investors are indispensable market participants for the oil futures and
derivatives markets
Financial investors (or speculators) have been playing a central role in the commodity futures
markets since the introduction of commodity futures. There are two primary functions of the
commodity futures: price discovery and risk management. Speculators take positive or negative
positions in the oil futures market with the purpose of making profit from price fluctuations. If a
speculator believes that new information justifies a higher valuation for crude oil, he takes a long
position in the oil futures market, and vice-versa. These trading activities, together with the trading
of commercial traders, serve to feed all of the available information into market prices. Hence, they
are the heart of oil futures markets.
A liquid and efficient market for long-dated oil futures is essential to meet the needs of risk
management from both oil producers and consumers. For example, a producer with an ongoing
exploration project in unconventional oil, like tar sands, may want to lock their profit by selling oil
futures with maturity greater than five years. But historically, trading of oil futures mainly focuses on
the front-month contract whereas the trading in long-dated oil contracts are very thin, which
severely limits the risk management role of the futures market.
The financialization process coincides with the rapid increase of the open interests in oil
futures, especially long-dated oil contracts. Figure 29 presents the open interests in crude oil
futures and adjusted option position for contracts with different maturities from 2000 to 2008.
Historically, the trading and open interests in long-dated oil futures are much thinner than those in
front-month contracts. For instance, the open interests in futures contracts expiring after three years
are very minor in 2000. Along with the rapid increase of the open interests in contracts expiring in
three months from 2000 to 2008, the open interests in contracts with longer maturities also grew
rapidly. The open interests in contracts with maturity between 12-36 months in 2008 are nearly twice
those in front-months contracts in 2000.
Open Interest refers to the total number of futures contracts that are not closed or delivered on a particular day. (See
Figure 1.17) The graph below shows that the trading and open interests in long-dated oil futures are much thinner
than those in front-month contracts.
Source: Büyükşahin et al. (2010).
Figure 29 – Open Interest in Crude Oil Futures and Adjusted Option Positions.
7.4.2 Under certain situations, the expectation may be easier to be distorted by market
Some of the statistical evidence discussed earlier suggests that oil prices may have deviated from the
fundamental values during certain periods, for example, the first half of 2008. In this section we
discuss several possible channels through which market expectation could be distorted.
Under efficient market hypothesis proposed by Fama (1970), the market is dominated by
rational investors and there is no room for the speculators to distort the prices of an asset because
the arbitrage behaviors of rational traders force the prices back to their fundamental levels. Noise
traders are defined as those who trade based on irrelevant information or sentiment. Although noise
traders are active in their trading activities, they have been largely neglected in the traditional
framework, because they tend to continuously lose money and their impacts on prices are limited.
This principle has been challenged by the recent behavioral finance literature, represented by
Black (1986), Shleifer and Summers (1990) and De Long, Shleifer, Summers and Waldmann (1990).
Under the new framework, some investors (noise traders) are not fully rational and their demand for
risky asset is affected by their beliefs or sentiments that are not fully justified by rational fundamental
news. Further, as arbitrage by rational traders is risky and therefore limited, asset prices can deviate
from their fair values for a long time. Some studies have shown that, on average, noise traders may
be more aggressive than arbitrageurs, because these traders are more optimistic and over-confident
and thus are likely to bear more risk. If higher risk is rewarded in the market, then noise traders can
earn higher expected returns on average, and hence as a group, they need not disappear from the
De Long et al. (1990) have shown that, if rational traders have short time horizons, due to
performance targets or reporting requirements, and if there are sufficiently many noise traders,
rational traders may choose to bet on continuation of the trend even though they realize the prices
deviated from fundamental levels. The 19992000 NASDAQ bubble appears to fit this description.
Brunnermeier and Nagel (2004) document the empirical evidence that hedge funds, who are
believed to be rational traders, did not exert a correcting force on stock prices during the technology
bubble. Instead, they rode the bubble by heavily investing in technology stocks. This does not seem
be the result of unawareness of the bubble: hedge funds captured the upturn, but, by reducing their
positions in stocks that were about to decline, avoided much of the downturn.
Inexperienced investors may play an important role in the formation of asset price
bubbles. Greenwood and Nagel (2009) use age as a proxy for managers’ investment experience and
find that around the peak of the technology bubble, mutual funds run by younger managers are
more heavily invested in technology stocks than their older colleagues. Furthermore, young
managers, but not old managers, exhibit trend-chasing behavior in their technology stock
investments. As a result, young managers increase their technology holdings during the run-up, and
decrease them during the downturn. Both results are in line with the behavior of inexperienced
investors in experimental asset markets. This result is relevant to the role of index investing in crude
oil markets. Most passive index investors, who mainly invest in conventional equity and bonds
markets, are relatively new to crude oil and commodities markets. They have just been attracted by
the recent robust performance of commodities in recent years and their investment in commodities
still remains a small proportion of their overall portfolio. It takes some time for them to obtain a
deeper understanding of the oil and commodities markets.
A new potential frontier for the role of financialization in oil prices may be the recent
approach developed by Singleton (2011). He argues that the existing economic models applied to
this issue are not capable of investigating the active role of financial activities. He proposed a
financial model based upon new developments in modern finance that allows him to incorporate a
more active role for financial investors. He attributes these results to two factors. Speculators’ ability
to arbitrage is limited by various financial market frictions on information and investment capital.
Additionally, heterogeneous investors with diverse beliefs can disagree on the interpretation of new
market conditions even when they share common public information. Using this new framework, he
finds significant empirical support that financial activities are likely to drive the oil price away from
their fundamental values. Investor flows influence excess returns from holding oil future contracts
of different maturities, after controlling for a number of other exogenous factors. He obtains
statistically significant and positive coefficients on variables for the lagged thirteen week change in
both imputed positions of index investors and managed-money spread positions. Further research
is needed, however, to confirm whether these results are robust with respect to data sources
(Buyuksahin 2011).
8.1 The Oil Price Explosion of 2004-2008
The oil price explosion of 2004-2008 reinforces the need for more thinking about the role of
producer and consumer expectations in the formation of crude oil prices. In the absence of storage,
one barrel of crude oil today is not exactly the same commodity as one barrel of crude oil five years
later. But long-run and short-run oil prices are linked together through the behavior of producers,
consumers, storage operators and speculators. With the anticipation of higher oil prices in the future,
producers may cut current oil production and leave it for later sale, and consumers may purchase
more oil now as precautionary demand. Both behaviors lower current supply and raise current
demand and thus drive up spot oil prices accordingly.
Oil inventories also act as a buffer for oil price shocks. With the price for oil futures
contract expiring one year later significantly higher than the spot price, it is profitable for someone
to buy oil today and store it with some storage costs, and sell it later. Such a trading strategy is
riskless arbitrage if they take a short position on a one-year contract at the same time. On the
contrary, when spot prices are higher than long-dated futures prices due to tight market conditions
or a short-run supply disruption, oil inventories will be released to the market. The extra oil supply
alleviates the market imbalance and therefore lowers spot prices. The arbitrage relationship among
the forward curves can be represented by the following equation11:
,=(1 + )
where r is the interest rate and represents the difference between storage cost and convenience
yields to physical inventories holders during time t to T.
Such a relationship is generally believed to describe well the price dynamics between spot
price, nearby futures and short-term or medium-term futures prices. Sometimes the oil futures
market exhibits the pattern of so-called “backwardation”, where nearby futures prices are higher
than the futures prices expiring later. One economic reason that explains this condition is a tight
market due to short-run supply shocks, and thus the convenience yields of holding physical
inventories dominate interest and storage costs. Instead, a “contango” market, in which nearby
futures prices are lower than futures price expiring later, can result from ample oil supply and less
incentives to hold physical inventories. According to Figure 30, from 1999 to 2004, backwardation is
11 The equation comes from the no arbitrage assumption. Consider the following portfolio: at time zero, the investor
borrows dollar and pays an interest rate, r. He also buys one unit of physical oil and then sells the one year forward
contract with price ,. After one year, the portfolio holder enjoys the convenience yield and delivers the physical oil
with the price F and then repays the debt (1+). Thus the net payoff to the portfolio holder is a constant number
,+-(1+). As the initial cost of this portfolio is zero, the assumption of no arbitrage implies that the net payoff
for such a strategy must be zero too. Otherwise, people will buy or sell an infinite number of such portfolios. Thus, we
have ,=(1+). Also see Geman (2005) Chapter two for details.
more common than contango, while after 2004, especially during 2005-2006, contango is more
frequently seen.
Figure 30 – Backwardation and Contango in 2000-2008 (Source: Büyükşahin et al., 2010).
The shapes of the oil forward curves have been changed with the financialization process.
Büyükşahin et al. (2008) identified and explained a structural change in the relation between crude
oil futures prices across contract maturities. As recently as 2001, near and long-dated futures were
priced as though traded in segmented markets. In 2002, however, the prices of one-year futures
started to move more in sync with the price of the nearby contract. Since mid-2004, the prices of
both the one-year-out and the two-year-out futures have been cointegrated with the nearby price.
In particular, these researchers utilize a unique dataset of individual traders to show that
increased market activity by commodity swap dealers, and by hedge funds and other financial traders,
has helped link crude oil futures prices at different maturities. Historically, the majority of oil futures
trading are through the front month contract. Since 2004, open interest and trading volume for
longer dated contracts have risen significantly, implying a greater liquidity and depth of the oil
market. The result is somewhat similar to that from examining the co-movements between the
convenience yields and oil price.
Through the channels discussed above, shocks to the long-run oil prices can be transmitted
to short-run prices and vice versa. This view highlights the role of expectation of market participants
about future oil demand and supply fundamentals in the formation of oil futures prices for both
front-end and long-dated contracts. Researchers are only beginning to understand these effects and
how to measure them empirically from the available data. But this issue will attract much more
attention in the future as oil-market analysts search for a framework that links short- and long-run
oil price movements.
8.2 Risk Premium Embedded in the Long-Dated Oil Futures Prices
Forecasting oil prices, especially the long-term oil prices, is important for both producers and
consumers. For instance, the major oil-producing countries, like Saudi Arabia, were under pressure
to gear their production and investment to the the oil price environment of $145 per barrel in 2008.
Like all rational investors and producers, their decision continued to depend more on the expected
long-run oil price, for example, in five to ten years. If the high oil prices are perceived by them to be
temporary and long-run oils are expected to move back to a lower level, there might be incentives
for them to produce closer to the maximum capacity. If oil prices are expected to stay high, there
may be incentives for additional capital investment to add to oil production capacity. Given the fact
that capacity investment is not reversible, the cost of making the wrong decision would be very high
as they are likely to experience lower oil prices if they produce more oil with the added capacity.
Beginning in March 1983 and extending through February 2007, Alquist and Kilian (2010)
generally confirmed the “no-change” rule for predicting the daily spot price for crude oil. No
change forecasts were preferred to the price of crude oil futures as well as to commercial survey-
based projections.
Although a biased signal and sometimes misleading, oil forward curves provide valuable real-
time information about the expected oil price in the futures by market participants. Market analysts
often cite the long-dated oil futures prices as the market forecast for long-term oil prices. Trading
activities by speculators also link the forward curve to the expected oil price in the future. From
finance theory, futures prices equal the expected spot prices in the future after accounting for the
risk premium, a component imbedded in oil futures prices to compensate the buys or sellers for
their exposure to the price risk. ,=() + risk premium
Risk premium reflects the aggregate risk preference of market participants. For commodities
like agricultural products, the risk premium is typically negative, that is, spot prices are lower than
expected prices in the future, as farmers could seek insurance by locking up the prices. Risk
premium could be affected by many factors, including macroeconomic conditions, market tightness,
funding cost of speculators, geopolitical risk and market sentiment. If we can find a good proxy for
risk premium, we should in principle derive a better forecast of oil prices in the future.
Pagano and Pisani (2009) document significant time-varying risk premium and use two such
proxies. One is the degree of capacity utilization in US manufacturing to represent the US business
cycle. The other is the oil inventory to represent oil market tightness. Manufacturing was running
well above the historical average, at almost 83 per cent. According to the forward curve, oil prices
were expected to decline to just over $20 by the following January. Risk-adjusted futures prices
according to the model of Pagano and Pisani (2009) were virtually indistinguishable from unadjusted
futures. This might suggest that there is no sizable risk premium imbedded in the oil forward curve.
Indeed, by January 1998 the oil price declined to $16.3 as was predicted by both market forward
curves and risk-adjusted prices from the model.
For the same reason, there are reasons to ask how large, if there is any, could be the risk
premium imbedded in the long-dated oil futures when oil price reached over $145 in July 2008 and
then when prices dropped sharply to $38 in December.
Figure 31 uses the forward curves for oil futures in two representative months to illustrate
the importance of risk premium in forecasting oil prices. As is shown in the upper panel, the oil spot
price was around $26 in January 1997 when demand was very high and capacity utilization in
manufacturing was running well above the historical average, at almost 83 per cent. According to the
forward curve, oil prices were expected to decline to just over $20 by the following January. Risk-
adjusted futures prices according to the model of Pagano and Pisani (2009) were virtually
indistinguishable from unadjusted futures. This might suggest that there is no sizable risk premium
imbedded in the oil forward curve. Indeed, by January 1998 the oil price declined to $16.3 as was
predicted by both market forward curves and risk-adjusted prices from the model.
The lower panel of Figure 31 presents a different story. Oil prices were stable at around $30
in September of 2003 and oil futures pointed to a decline in oil price to just below $26 in the
following 12 months. During that period of time, the recovery out of the recession in 2001 was not
yet firmly established and the capacity utilization index was still relatively low, at around 73 per cent.
Although the oil prices were expected to be as high as $38 by September 2004, as suggested by the
risk-adjusted prices, the market participants would only like to buy 12-month oil futures at $26. The
buyers need the risk premium as high as $12 to compensate their exposure of the downside risk.
This figure presents the results of risk-adjusting method by : Pagano and Pisani (2009),
Source: Pagano and Pisani (2009), p22
Figure 31Oil Price Forecasts and Realized Spot Prices on Two Dates.
Most research and studies on the long-run oil price paths have focused on stable trends over
decades without trying to anticipate cyclical oil price behavior. Demand factors are captured in
trends conditioned by the best available estimates for the price and income elasticities based upon
historical evidence and judgment. Supply factors are introduced through assessments of the costs
and availability of oil resources as well as the behavior of major oil exporters.
Several important challenges must be overcome before long-run oil price paths can be better
understood. First, there is no credible, integrated and comprehensive theory about the market
behavior of the OPEC members. Without this basic understanding, it is unrealistic to expect that oil
price behavior can be understood, even if analysts understand oil demand and non-OPEC supply
trends adequately. Although many experts recognize the importance of geopolitical considerations,
they have been unable to develop a reliable, empirical approach for incorporating these issues.
Second, technical change or technical progress tends to shift the supply and demand trends
gradually but significantly. This issue deserves much greater attention in future analysis of oil market
trends, because these irreversible shifts will have serious implications for current long-term
investments in capacity, production and end-use equipment. In the early 1980s, seismic and
horizontal drilling improved oil supply prospects outside the Arabian Gulf. After the oil price
shocks of the 1970s, government policy and the automobile industry transformed the vehicle stock
to increase miles per gallon. Both effects continued strongly, despite the decline in crude oil prices
over the 1980s and into the 1990s. Today, potential supply shifts may include the promising
developments in shale gas resources, which is changing the energy balance picture in the U.S and
might potentially make similar changes worldwide. Other technical developments in the
transportation sector include ethanol additives and hybrid vehicles. It seems unrealistic to expect
these possible substitutes to remain on the back burner indefinitely, if future oil prices continue to
track upward.
And third, the oil price explosion of 2004-2008 reinforces the need for more thinking about
the role of producer and consumer expectations in the formation of crude oil prices. Expectations
about oil market conditions are contingent upon not only current levels or recent past trends but
also judgments about future events. However, analyzing the short-run dynamics of oil prices is very
complicated. It requires a careful separation of the fundamental market conditions from the
increasing role played by financial markets to disentangle the impacts of such factors as inelastic
demand and supply, risk premium, market sentiment and bandwagon speculation. As the world
economy begins to recover from the 2008 financial collapse, these expectations will change quickly
with many possible surprises. These developments make it all the more challenging to undertake
production, quota or investment policies for stabilizing oil prices.
The main role for oil-related financial instruments is to communicate expectations as
efficiently as possible. Although it is difficult to document a long-standing oil financial bubble from
recent events, the 2004-2008 period has provided the impetus for perhaps a much-needed closer
supervision and review of commodity futures markets. Eliminating or severely restricting these
financial effects through new forms of aggressive regulation, however, may block the pathway to a
more efficiently operating oil market where expectations are easily communicated to all market
This paper is based on research project supported fully by King Abdullah Petroleum Studies and
Research Center (KAPSARC), Riyadh, Saudi Arabia. We would like to acknowledge the careful
comments and review provided by Majid Moneef, Bassam Fattouh, James Hamilton, James Smith,
Coby van der Linde, and Ali Aissaoui. We also benefitted from the valuable advice from and
discussion with John P. Weyant, Stephen P.A. Brown, James L. Sweeney and participants at the Oil
Metric Forum workshop in Washington, D.C. on September 16-17, 2010, and the National Energy
Policy Institute Conference on OPEC at 50: Its Past, Present and Future in a Carbon-Constrained
World in Tulsa, OK on March 23, 2011. Al-Fattah would like to thank Walid Matar and Jiangjiang
Pan for their support. All responsibility for the contents of this paper belongs to the principal
authors, and none of the views and conclusions can be attributed to any of the above individuals or
the King Abdullah Petroleum Studies and Research Center (KAPSARC).
Adelman, M. A. (1964). Efficiency and Resource Use in Crude Petroleum. Southern Economic Journal,
31, 101116.
Adelman, M. A. (1980). The Clumsy Cartel. The Energy Journal, 1(1), 43-54.
Adelman, M. A. (1982). OPEC as a Cartel. in OPEC Behavior and World Oil Prices, J. Griffin and D.
Teece, eds., London: Allen & Unwin, 37-57.
Adelman, M. A. (1986). Oil Producing Countries’ Discount Rates. Resources and Energy, 8, 309-329.
Adelman, M. A. (1990). Mineral Depletion, with Special Reference to Petroleum. The Review of
Economics and Statistics, 72(1), 1-10.
Adelman, M. A., and Manoj Shani. (1989). Oil development-operating cost estimates, 1955-1985.
Energy Economics, 11(1), 2-10.
Adeyemi, Olutomi I., and Lester C. Hunt. (2007). Modelling OECD industrial energy demand:
Asymmetric price responses and energy-saving technical change. Energy Economics, 29(4), 693-709.
Adeyemi, Olutomi I., David C. Broadstock, Mona Chitnis, Lester C. Hunt, and Guy Judge. (2010).
Asymmetric price responses and the underlying energy demand trend: Are they substi tutes or
complements? Evidence from modelling OECD aggregate energy demand. Energy Economics, 32
(5), 1157-1164.
Aguilera, Roberto F., Roderick G. Eggert, Gustavo Lagos C.C., and John E. Tilton. (2009).
Depletion and the Future Availability of Petroleum Resources. The Energy Journal, 30(1), 141-174.
Alhajji, A.F, and David Huettner. (2000). OPEC and World Crude Oil Markets from 1973 to 1994:
Cartel, Oligopoly, or Competitive? The Energy Journal, 21(3), 31-60.
Al-Qahtani, Ayed, Edward Balistreri and Carol Dahl. (2008). Literature Review on Oil Market Modeling
and OPEC’s Behavior. Division of Economics and Business, Colorado School of Mines, Golden
Al-Yousef, Nourah A. (1998). Modeling Saudi Arabia Behavior in the World Oil Market 1976- 1996.
Department of Economics, University of Surrey. SEEDS 93.
Alquist, Ron, and Lutz Kilian. (2010). What do we learn from the price of crude oil futures? Journal of
Applied Econometrics, 25(4), 539-573.
Almoguera, Pedro A. and Ana Maria Herrera. (2007). Testing for the Cartel in OPEC: Noncooperative
Collusion or Just Noncooperative? mimeo, Michigan State University.
Baldwin, Nick, Richard Prosser. (1988). World oil market simulation. Energy Economics, 10(3), 185-
Bartis, James T, Tom LaTourrette, Lloyd Dixon, D.J. Peterson and Gary Cecchine. (2005). Oil Shale
Development in the United States: Prospects and Policy Issues. Prepared for the National Energy
Technology Laboratory of the United States Department of Energy, The RAND Corporation.
Baumeister, Christiane and Peersman, Gert. (2010). Sources of the Volatility Puzzle in the Crude Oil
Market. Working Papers of Faculty of Economics and Business Administration, Ghent
University, Belgium, Ghent.
Beenstock, Michael, and Patrick Wilcocks. (1981). Energy Consumption and Economic Activity in
Industrialized Countries: The Dynamic Aggregate Time Series Relationship. Energy Economics, 3(4),
Beenstock, Michael, and Patrick Wilcocks. (1983). Energy and Economic Activity: A Reply to
Kouris. Energy Economics, 5(3), 212-212.
Bentzen, Jan, and Tom Engsted. (1996). On the Estimation of Short- and Long-Run Elasticities in
U. S. Petroleum Consumption: Comment. Southern Economic Journal (Southern Economic Association),
62(3), 783-787.
Black, Fischer. (1986). Noise. Journal of Finance, 41(3), 529-43.
Brennan, M. J. (1958). The Supply of Storage. The American Economic Review, 47, 50-72.
Brown, S., R. Virmani and R. Alm. (2008). Crude Awakening: Behind the Surge in Oil Prices, FRB
Dallas Economic Letter, 3(5), 1–8.
Brunnermeier, Markus K. (2009). Bubbles. Entry in The New Palgrave Dictionary of Economics, edited by
Steven Durlauf and Lawrence Blume, 2nd edition.
Brunnermeier, Markus K., and Stefan Nagel. (2004). Hedge Funds and the Technology Bubble.
Journal of Finance, 59(5), 2013-2040.
Buyuksahin, Bahattin. (2011). Commodity Index Funds: The New Whipping Boys? Online blog,
September 27, 2011, available at
Buyuksahin, Bahattin, Michael S. Haigh, Jeffrey H. Harris, James A. Overdahl and Michel A. Robe.
(2008). Fundamentals, Trader Activity and Derivative Pricing. EFA 2009 Bergen Meetings Paper.
Buyuksahin, Bahattin and Jeffrey H. Harris. (2009). The Role of Speculators in the Crude Oil Futures
Market, July 16, Available at SSRN:
Buyuksahin, Bahattin, Celso Brunetti, Michel A. Robe, and Kirsten R. Soneson. (2010). Do OPEC
Members Know Something the Market Doesnt? Fair Price Pronouncements and the Market
Price of Crude Oil. SSRN eLibrary (SSRN).
Buyuksahin, Bahattin and Michel A. Robe. (2010). Speculators, Commodities, and Cross-Market Linkages.
working paper, American University.
Cairns, Robert D. (1990). The Economics of Exploration for Non-renewable Resources. Journal of
Economic Surveys, 4(4), 361-95.
CFTC. (2008). CFTC Staff Report on Commodity Swap Dealers and Index Traders with Commission
Recommendations. CFTC.
Cleveland, Cutter J., and Robert K. Kaufmann. (1991). Forecasting Ultimate Oil Recovery and Its
Rate of Production: Incorporating Economic Forces into the Models of M. King Hubbert. The
Energy Journal, 12(2), 17-46.
Cuddington, John T., and Diana L. Moss. (2001). Technological Change, Depletion, and the U.S.
Petroleum Industry. American Economic Review, 91(4), 1135-1148.
Dahl, Carol, and Mine Yucel. (1991). Testing Alternative Hypotheses of Oil Producer Behavior. The
Energy Journal, 12(4), 117-138.
Dahl, Carol, and Thomas Sterner. (1991). Analysing Gasoline Demand Elasticities: A Survey. Energy
Economics, 13(3), 203-210.
Dargay, Joyce, and Dermot Gately. (1995). The Imperfect Price Reversibility of Non-Transport Oil
Demand in the OECD. Energy Economics, 17(1), 59-71.
Dargay, Joyce, and Dermot Gately. (2010). World Oil Demand's Shift Toward Faster Growing And
Less Price-Responsive Products And Regions. Energy Policy, 38(10), 6261-6277.
Dargay, Joyce, Dermot Gately, and Martin Sommer. (2007). Vehicle Ownership and Income
Growth, Worldwide: 1960-2030. The Energy Journal, 28(4), 143-170.
Davidson, Paul. (2008). Crude Oil Prices: ‘Market Fundamentals’ or Speculation? Challenge, 51(4),
De Long, J Bradford, Andrei Shleifer, Lawrence H. Summers, and Robert J. Waldmann. (1990).
Noise Trader Risk in Financial Markets. Journal of Political Economy, 98(4), 703-38.
Dees, Stephane, Pavlos Karadeloglou, Robert K. Kaufmann, and Marcelo Sanchez. (2007).
Modelling the World Oil Market: Assessment of a Quarterly Econometric Model. Energy Policy,
35(1), 178-191.
Dvir, Eyal, and Kenneth S. Rogoff. (2009). Three Epochs of Oil. NBER Working Papers, National
Bureau of Economic Research, Inc.
Einloth, James T. (2009). Speculation and Recent Volatility in the Price of Oil. SSRN eLibrary
Eller, Stacy, Peter Hartley, and Kenneth Medlock. (2010). Empirical Evidence on the Operational
Efficiency of National Oil Companies. Empirical Economics, 1-21.
Energy Modeling Forum. (2010). Forum on Oil Price Determinants. Stanford, CA, Energy Modeling
Energy Modeling Forum. (1982). World Oil. Stanford University, Stanford, CA.
Etula, Erkko. (2009). Broker-dealer risk appetite and commodity returns. Staff Reports, Federal Reserve
Bank of New York.
Fama, Eugene F.(1970). Efficient Capital Markets: A Review of Theory and Empirical Work. Journal
of Finance, 25(2), 383-417.
Farrell, A. E. and A. R. Brandt. (2006). Risks of the Oil Transition. Environmental Research Letters, 1,
Fattouh, Bassam. (2007). The Drivers of Oil Prices: The Usefulness and Limitations of Non-Structural Model,
the DemandSupply Framework and Informal Approaches. Oxford Institute of Energy Studies, WPM32
Fattouh, Bassam. (2010a). Oil Market Dynamics through the Lens of the 2002-2009 Price Cycle. Oxford
Institute for Energy Studies.
Fattouh, Bassam. (2010b). The dynamics of crude oil price differentials. Energy Economics, 32(2), 334-
Fischer, Dietrich, Dermot Gately, and John F. Kyle. (1975). The Prospects for OPEC: A Critical
Survey of Models of the World Oil Market. Journal of Development Economics, 2(4), 363-386.
Fisher, Franklin M., Carl Kaysen. (1962). A Study in Econometrics: The Demand for Electricity in the United
States. North-Holland Publishing Co, Amsterdam.
Galli, Rossana. (1998). The Relationship Between Energy Intensity and Income Levels: Forecasting
Long-term Energy Demand in Asian Emerging Countries. The Energy Journal, 19(4), 85-106.
Gately, Dermot. (2004). OPEC's Incentives for Faster Output Growth. The Energy Journal, 25(2), 75-
Gately, Dermot. (2007). What Oil Export Levels Should We Expect From OPEC? The Energy Journal,
28(2), 151-174.
Gately, Dermot, John F. Kyle, and Dietrich Fischer. (1977). Strategies for OPEC's Pricing
Decisions. European Economic Review, 10(2), 209-230.
Gately, Dermot and Hillard G. Huntington. (2002), The Asymmetric Effects of Changes in Price
and Income on Energy and Oil Demand. The Energy Journal, 23(1), 19-56.
Geman, Héylette. (2005). Commodities and Commodity Derivatives - Modeling and Pricing for Agriculturals,
Metals and Energy, Wiley.
Gilbert, Christopher L. (2010). Speculative Influences On Commodity Futures Prices 20062008. Working
Gillingham, Kenneth. (2011). How Do Consumers Respond to Gasoline Price Shocks? Heterogeneity in Vehicle
Choice and Driving Behavior. Working paper.
Glomsrød, Solveig and Petter Osmundsen. (2005). Petroleum industry regulation within stable states.
Aldershot: Ashgate.
Goodwin, Phil, Joyce Dargay and Mark Hanly. (2004). Elasticities of Road Traffic and Fuel
Consumption with Respect to Price and Income: A Review. Transport Reviews: A Transnational
Transdisciplinary Journal, 24(3), 275-292.
Gorton, Gary, and K. Geert Rouwenhorst. (2004). Facts and Fantasies about Commodity Futures. NBER
Working Papers, National Bureau of Economic Research, Inc.
Graham, Daniel J. and Stephen Glaister. (2004). Road traffic demand elasticity estimates: A review.
Transport Reviews: A Transnational Transdisciplinary Journal, 24(3), 261274.
Gracia, Eduard. (2006). Bubbling Crude: Oil Price Speculation and Interest Rates. e-journal of petroleum
management and economics, 1(1).
Granger. C. W. J. (1969). Investigating Causal Relations by Econometric Models and Cross-Spectral
Methods. Econometrica, 37(3), 424-38.
Greenwood, Robin, and Stefan Nagel. (2009). Inexperienced investors and bubbles. Journal of
Financial Economics, 93(2), 239-258.
Griffin, James M. (1985). OPEC Behavior: A Test of Alternative Hypotheses. American Economic
Review, 75(5), 954-63.
Griffin, James M. (1992). OPEC and World Oil Prices: Is the Genie Back in the Bottle? Energy
Studies Review, 4(1).
Griffin, James M, and William S Neilson. (1994). The 1985-86 Oil Price Collapse and Afterwards:
What Does Game Theory Add? Economic Inquiry, 32(4), 543-61.
Griffin, James M., and Craig T. Schulman. (2005). Price Asymmetry in Energy Demand Models: A
Proxy for Energy-Saving Technical Change? The Energy Journal, 26(2), 1-22.
Gulen, S. Gurcan. (1996). Is OPEC a Cartel? Evidence from Cointegration and Causality Tests. The
Energy Journal, 17(2), 43-58.
Hamilton, James D. (2009a). Causes and Consequences of the Oil Shock of 2007-08. Brookings Papers
on Economic Activity, Spring 2009, 215-259.
Hamilton, James D. (2009b). Understanding Crude Oil Prices. The Energy Journal, 30(2), 179-206.
Hartley, Peter, and Kenneth B. Medlock. (2008). A Model of the Operation and Development of a
National Oil Company. Energy Economics, 30(5), 2459-2485.
Headey, Derek, and Shenggen Fan. (2008). Anatomy of a Crisis: The Causes and Consequences of Surging
Food Prices. IFPRI discussion papers, International Food Policy Research Institute (IFPRI).
Hicks, Bruce, and Lutz Kilian. (2009). Did Unexpectedly Strong Economic Growth Cause the Oil Price Shock
of 2003-2008? CEPR Discussion Papers, C.E.P.R. Discussion Papers.
Hnyilicza, Esteban, and Robert S. Pindyck. (1976). Pricing Policies for a Two-Part Exhaustible
Resource Cartel: The Case of OPEC. European Economic Review, 8(2), 139-154.
Horn, Manfred. (2004). OPEC's optimal crude oil price. Energy Policy, 32(2), 269-280.
Hosier, Richard H. (2004). Energy Ladder in Developing Nations. In Encyclopedia of Energy, edited by
Cutler J. Cleveland, Elsevier, New York, 423-435.
Hotelling, Harold. (1931). The Economics of Exhaustible Resources. Journal of Political Economy, 39,
Hubbert, M. King. (1956). Nuclear Energy and the Fossil Fuel. American Petroleum Institute, Drilling
and Production Practice.
Hubbert, M. King. (1962). Energy Resources, Publication 1000-d. National Academy of Sciences-
National Research Council, Washington, D.C.
Hughes, Jonathan E., Christopher R. Knittel, and Daniel Sperling. (2008). Evidence of a Shift in the
Short-Run Price Elasticity of Gasoline Demand. The Energy Journal, 29(1), 113-134.
Hunt, Lester C., and Yasushi Ninomiya. (2003). Unravelling Trends and Seasonality: A Structural
Time Series Analysis of Transport Oil Demand in the UK and Japan. The Energy Journal, 24(3), 63-
Huntington, Hillard G. (1994). Oil Price Forecasting in the 1980s: What Went Wrong? The Energy
Journal, 15(2), 1-22.
Huntington, Hillard G. (2006). A Note on Price Asymmetry as Induced Technical Change. The
Energy Journal, 27(3), 1-8.
Huntington, Hillard G. (2010a). Short- and Long-Run Adjustments in U.S. Petroleum Consumption.
Energy Economics, 32(1) 63-72.
Huntington, Hillard G. (2010b). Oil Demand and Technical Progress. Applied Economics Letters, DOI:
10.1080/ 13504850903299636
Interagency Task Force on Commodity Markets. (2008). Interim Report on Crude Oil. CFTC,
Washington, D.C..
International Energy Agency. (2010). End-Use Petroleum Product Prices and Average Crude Oil Import
Costs, Paris, France.
Irwin, Scott H., Dwight R. Sanders, and Robert P. Merrin. (2009). Devil or Angel? The Role of
Speculation in the Recent Commodity Price Boom (and Bust). Journal of Agricultural and Applied
Economics, 41(2).
Irwin, Scott H., Dwight R. Sanders. (2010). The impact of index swap funds on commodity futures
markets. Technical report, OECD Food, Agriculture and Fisheries Working Papers No. 27.
Johansson, Daniel J.A., Christian Azar, Kristian Lindgren and Tobias A. Persson. (2009). OPEC
Strategies and Oil Rent in a Climate Conscious World. The Energy Journal, 30(3), 23-50.
Johany, Ali D. (1979). OPEC and the price of oil: cartelization or alteration of property rights.
Journal of Energy and Development, 5, 7280.
Johnson, David, Daniel Johnston, and Tony Rogers. (2008). International Petroleum Taxation.
Independent Petroleum Association of America. Available at:
Jones, Clifton T. (1990). OPEC Behavior Under Falling Prices: Implications for Cartel Stability, The
Energy Journal, 11(3), 117-129.
Jones, Clifton T. (1993). A Single-Equation Study of U. S. Petroleum Consumption: The Role of
Model Specification. Southern Economic Journal (Southern Economic Association), 59(4), 687-700.
Kaufmann, Robert K., and Cutler J. Cleveland. (2001). Oil Production in the Lower 48 States:
Economic, Geological, and Institutional Determinants. The Energy Journal, 22(1), 27-50.
Kemp, Alexander G., and D. Crichton. (1979). Effects of Changes in UK North Sea Oil Taxation.
Energy Economics, 1(4), 224-240.
Kilian, Lutz. (2009). Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks
in the Crude Oil Market. American Economic Review, 99(3), 1053-69.
Kilian, Lutz. (2010). Explaining Fluctuations in Gasoline Prices: A Joint Model of the Global Crude
Oil Market and the U.S. Retail Gasoline Market. The Energy Journal, 31(2), 87-112.
Kilian, Lutz, and Dan Murphy. (2010). The Role of Inventories and Speculative Trading in the Global Market
for Crude Oil. CEPR Discussion Papers.
King Abdullah Petroleum Studies and Research Center. (2011). Inventory of Research and Analysis on
Overall Oil Price Determinants, Riyadh, Saudi Arabia.
Knittel, Christopher R. (2009). Automobiles on Steroids: Product Attribute Trade-Offs and Technological
Progress in the Automobile Sector. NBER Working Papers 15162, National Bureau of Economic
Kouris, George. (1983). Energy Consumption and Economic Activity in Industrialized Economies--
a Note. Energy Economics, 5(3), 207-212.
Krautkraemer, Jeffrey A. (1998). Nonrenewable Resource Scarcity. Journal of Economic Literature, 36(4),
Krichene, Noureddine. (2002). World crude oil and natural gas: a demand and supply model. Energy
Economics, 24(6), 557-576
Krugman, Paul. (2008). The Oil Non-bubble, New York Times.
Libecap, Gary D. and James L. Smith. (2004). Political Constraints on Government Cartelization:
The Case of Oil Production Regulation in Texas and Saudi Arabia. Chapter 8 in How Cartels22
Endure and How They Fail: Studies in Industrial Collusion, ed. by Peter Z. Grossman, Cheltenham,
Edward Elgar.
Litzenberger, Robert H and Nir Rabinowitz. (1995). Backwardation in Oil Futures Markets: Theory
and Empirical Evidence. Journal of Finance, 50(5), 1517-45.
Liu, Peng and Ke Tang. (2010). Bubbles in the Commodity Asset Class: Detection and Sources.
Livernois, John R, and Russell S Uhler. (1987). Extraction Costs and the Economics of
Nonrenewable Resources. Journal of Political Economy, 95(1), 195-203.
Lynch, Michael C. (2002). Forecasting Oil Supply: Theory and Practice. The Quarterly Review of
Economics and Finance, 42(2), 373-389.
MacAvoy, Paul W. (1982). Crude Oil Prices: As Determined by OPEC and Market Fundamentals.
Cambridge: Ballinger.
Marcel, Valerie. (2006). Oil Titans: National Oil Companies in the Middle East, Chatham House, London.
Masters, Michael W. (2008). Testimony before the Committee on Homeland Security and
Governmental Affairs United States Senate.
Masters, Michael W. and Adam K. White. (2009). The 2008 Commodities Bubble: Assessing the Damage to
the United States and Its Citizens. Special report.
Maugeri, Leonardo. (2009). Understanding Oil Price Behavior through an Analysis of a Crisis. Review
of Environmental Economics and Policy, 147-166.
Mead, Walter J. (1979). The Performance of Government in Energy Regulations. American Economic
Review, 69(2), 352-56.
Medlock, Kenneth B. and Amy Myers Jaffe. (2009). Who is in the oil futures market and how has it
changed. The James A. Baker Institute for Public Policy of Rice University.
Medlock, Kenneth B., and Ronald Soligo. (2001). Economic Development and End-Use Energy
Demand. The Energy Journal, 22(2), 77-106.
Moran, Theodore. (1982). Modeling OPEC Behavior: Economic and Political Alternatives, in J. M.
Griffin and D. J. Teece, eds., OPEC Behavior and World Oil Prices, London: Allen & Unwin.
Nordhaus, William D. (2009). The Economics of an Integrated World Oil Market. Paper presented
at the International Energy Workshop, Venice, Italy.
Pagano, Patrizio, and Massimiliano Pisani. (2009). Risk-Adjusted Forecasts of Oil Prices. The B.E.
Journal of Macroeconomics, 9(1).
Parsons, John E. (2009). Black Gold & Fool’s Gold: Speculation in the Oil Futures Market. Working
Papers, Massachusetts Institute of Technology, Center for Energy and Environmental Policy
Phillips, Peter Charles Bonest and Jun Yu. (2009). Dating the Timeline of Financial Bubbles During
the Subprime Crisis. Research Collection School of Economics. Paper 1153.
Pierru, Axel, and Denis Babusiaux. (2010). Speculation without Oil Stockpiling as a Signature: A Dynamic
Perspective. Working Papers, Massachusetts Institute of Technology, Center for Energy and
Environmental Policy Research.
Pindyck, Robert S. (1977). Cartel Pricing and the Structure of the World Bauxite Market. Bell Journal
of Economics, 8(2), 343-360.
Pindyck, Robert S. (1978a). Gains to Producers from the Cartelization of Exhaustible Resources. The
Review of Economics and Statistics, 60(2), 238-51.
Pindyck, Robert S. (1978b). The Optimal Exploration and Production of Nonrenewable Resources.
Journal of Political Economy, 86(5), 841-61.
Pindyck, Robert S. (1979). Interfuel Substitution and the Industrial Demand for Energy: An
International Comparison. The Review of Economics and Statistics, 61(2), 169-79.
Pindyck, Robert S. (1980). Uncertainty and Exhaustible Resource Markets. Journal of Political Economy,
88(6), 1203-25.
Pindyck, Robert S. (1993). The Present Value Model of Rational Commodity Pricing. Economic
Journal, 103(418), 511-30.
Pindyck, Robert S. (1999). The Long-Run Evolutions of Energy Prices. The Energy Journal, 20(2), 1-
Pirog, Robert. (2007). The Role of National Oil Companies in the International Oil Market, Congressional
Research Service, Washington, DC.
Plourde, Andre. (2009). Oil Sands Royalties and Taxes in Alberta: An Assessment of Key
Developments since the mid-1990s. The Energy Journal, 30(1), 111-140.
Powell, Stephen G. (1990). The Target Capacity-Utilization Model of OPEC and the Dynamics of
the World Oil Market. The Energy Journal, 11(1), 27-64.
Ryan, David L., and Andre Plourde. (2002). Smaller and Smaller? The Price Responsiveness of
Nontransport Oil Demand. The Quarterly Review of Economics and Finance, 42(2), 285-317.
Salant, Stephen W. (1976). Exhaustible Resources and Industrial Structure: A Nash-Cournot
Approach to the World Oil Market. Journal of Political Economy, 84(5), 1079-93.
Sanders, Dwight R. and Scott H. Irwin. (2010). A Speculative Bubble in Commodity Futures Prices?
Cross-Sectional Evidence. Agricultural Economics, 41(1), 25-32.
Sanders, Dwight R. and Scott H. Irwin. (2011). The Impact of Index Funds on U.S. Grain Futures
Prices. Journal of Alternative Investments (Forthcoming).
Schipper, Lee, Ipsita Banerjee and Wei-Shiuen Ng. (2009). Carbon Dioxide Emissions from Land
Transport in India. Journal of the Transportation Research Board, 2114(1), 28-37.
Shleifer, Andrei, and Lawrence H. Summers. (1990). The Noise Trader Approach to Finance. Journal
of Economic Perspectives, 4(2) 19-33.
Singleton, Kenneth J. (2011). Investor Flows and the 2008 Boom/Bust in Oil Prices. Graduate School of
Business, Stanford University.
Slade, Margaret E. (1982). Trends in Natural-Resource Commodity Prices: An Analysis of the Time
Domain. Journal of Environmental Economics and Management, 9(2), 122-137.
Small, Kenneth A., and Kurt Van Dender. (2007). Fuel Efficiency and Motor Vehicle Travel: The
Declining Rebound Effect. The Energy Journal, 28(1), 25-52.
Smith, James L. (2005). Inscrutable OPEC? Behavioral Tests of the Cartel Hypothesis. The Energy
Journal, 26(1). 51-82.
Smith, James L. (2009). World Oil: Market or Mayhem? Journal of Economic Perspectives, 23(3), 145-64.
Smith, James L. (2012). On the Portents of Peak Oil (And Other Indicators of Resource Scarcity.
Energy Policy (forthcoming).
Smith, James L and James L. Paddock. (1984). Regional Modelling of Oil Discovery and
Production. Energy Economics, 6(1), 5-13.
Stoll, Hans R., and Robert E. Whaley. (2010). Commodity Index Investing: Speculation or
Diversification? SSRN eLibrary (SSRN).
Tang, Ke, and Wei Xiong. (2010). Index Investment and Financialization of Commodities. NBER
Working Papers, National Bureau of Economic Research, Inc.
Teece, David J. (1982). OPEC Behavior: An Alternative View in OPEC Behavior and World Oil Prices,
J. Griffin and D. Teece, eds., London: Allen & Unwin.
Telser, Lester G. (1958). Futures Trading and the Storage of Cotton and Wheat. Journal of Political
Economy, 66, 233.
Till, Hilary. (2009). Has There Been Excessive Speculation in the US Oil Futures Markets? What Can We
(Carefully) Conclude from New CFTC Data? EDHEC-Risk Publication.
United States Energy Information Administration, (2010). International Energy Outlook. #:DOE/EIA-
0484(2010), Washington, D.C..
United States Geological Survey. (2005). World Petroleum Assessment 2000: New Estimates of Undiscovered
Oil and Natural Gas, Natural Gas Liquids, Including Reserve Growth, Outside the United States, Reston,
United States Senate, (2006). The Role of Market Speculation in Rising Oil and Gas Prices: A Need to Put the
Cop Back on the Beat. Staff report prepared by the permanent subcommittee on investigations of
the committee on Homeland Security and Governmental Affairs, Washington, D.C..
Van Benthem, Arthur, and Mattia Romani. (2009). Fuelling Growth: What Drives Energy Demand
in Developing Countries? The Energy Journal, 30(3), 91-114.
Verleger, Philip K Jr. (2009). Anatomy of the 10-Year Cycle in Crude Oil Prices, University of
Calgary, Calgary, Alberta, Canada.
Walker, I.O., and Franz Wirl. (1993). Irreversible Price-Induced Efficiency Improvements: Theory
and Empirical Application to Road Transportation. The Energy Journal, 14(4), 183-205.
Watkins, G. Campbell. (2002). Characteristics of North Sea Oil Reserve Appreciation. In Oil and
The Economy: Recent Developments in Historical Perspective. Quarterly Review of Economics and
Finance, special issue ed. by James L. Smith, 47(2), 335-372.
Wirl, Franz, and Azra Kujundzic. (2004). The Impact of OPEC Conference Outcomes on World
Oil Prices 1984-2001. The Energy Journal, 25(1), 45-62.
Wolf, Christian. (2009). Does Ownership Matter? The Performance and Efficiency of State Oil vs.
Private Oil (1987-2006). Energy Policy, 37(7), 2642-2652.
... As a country's economic growth is enhanced and incomes increase, oil consumption eventually begins to increase. This in turn creates a causal relationship between economic growth and fluctuations in crude oil prices (Huntington et al., 2012). Hanabusa (2009) conducted a study where he utilized an AR-EGARCH model to study the relationship between economic growth and oil prices between 2000 and 2008 in Japan. ...
... Speculation within the oil market tends to play a role in influencing oil price movements as participants in the oil market are divided into hedgers who seek to minimize their exposure to risk and speculators who seek to make a profit as a result of the movement of oil prices (Huntington et al., 2012). Xiao and Wang (2022) implemented a study in which they aimed to determine the relationship between uncertainty and oil price fluctuations by utilizing the macroeconomic uncertainty (MU) index as an indicator. ...
... However, excess supply or surplus always leads to increased crude oil inventories. (Huntington et al., 2012) Two types of oil storage are recommended to exist: public storage by the government, and private storage by the private sector. Crude storage by the government is important to reduce oil supply disruptions. ...
The beginning of the new century was marked with another petroleum boom and bust cycle. Oil prices were hovering around $18-20/bbl through most of the 1990s after which crude prices collapsed to $10/bbl in 1998 and 1999. Soon thereafter oil prices began a steady and, at times, sharp rise on the way to $147/bbl in July 2008. This climb was followed by an abrupt decline after the onset of the global “Lehman” economic crises in September 2008 driving down the crude oil price to as low as $32/bbl in December 2008. After a relatively swift recovery, another oil shock “market share” took place in 2014-2016; average oil prices plunged from $108/bbl in the second quarter of 2014 to $30/bbl in the first quarter of 2016. Brent kicked off 2018 with average oil prices of $69/bbl in January toggling around the $85/bbl during October 2018; since then, however, oil prices were dwindling from $80/bbl to $51/bbl by 2018 year end. The year of 2019 started with a fluctuating Brent oil prices around the 55-65$/bbl range. The rapid increase in the oil price and its sudden and dramatic decline raises a fundamental question about the oil industry: Why is it so difficult to accurately predict the price of oil? Supply-demand balance, economic growth, oil inventories, and spare capacity are market fundamentals that drive oil prices and market dynamics. Market financialization, resources availability, technology advancements, and geopolitical events are also important drivers of oil price movements. Collaborative efforts should be geared towards: an acceptable and reasonable level of oil prices for the benefits of oil producers and consumers alike; meeting the future oil demand and availing adequate spare capacity to the market; and incentivizing upstream capital investment. Reasonable predictions of oil prices require a reliable and consistent data, rigorous advanced analytical methods, intelligent forecasting tools, and a better understanding of the influential factors impacting the oil prices and oil market conditions.
... Econometric models that predict oil prices, oil production, and OPEC behaviour are not the focus of this paper; see Al-Qahtani et al. (2008) and Huntington et al. (2012) for a survey on these subjects. Dixit (1984) discusses the literature of trade policies for imperfectly competitive industries without externalities, which is also not discussed in my paper. ...
Full-text available
Climate change is a challenging and urgent problem. Common policies for combating climate change (e.g., taxes or quotas) assume perfectly competitive markets to achieve the First Best, although in reality many pollution-intensive markets are not perfectly competitive. This paper summarizes the literature that examines the most common climate policies when the market is not perfectly competitive. The literature that shows that firms with market power act strategically to manipulate prices to their advantage. When tradable permits are imposed, dominant firms manipulate both permit and output prices. When a world carbon tax is imposed, OPEC may increase or decrease the oil price depending on whether there is threatening entries. Even before a world carbon tax is imposed, OPEC strategically increases the oil price to prevent it. When a global emission target is chosen by a region unilaterally, OPEC manipulates prices to go up, which drastically decreases leakages and shifts abatement costs from the regulated regions to the unregulated regions.
... This paper is based on research supported by King Abdullah Petroleum Studies and Research Center (KAPSARC) and contained in Huntington et al (2012Huntington et al ( , 2013, which provides a much more extensive set of references than we could produce here. We would like to acknowledge the careful comments and review on our previous work by Bassam Fattouh, James Hamilton, James Smith, Fred Joutz, Majid Moneef, and Awwad Al-Harthi. ...
Full-text available
The complexity of the world oil market has increased dramatically in recent years and new approaches are needed to understand, model, and forecast oil prices today. In addition to the commencement of the financialization era in oil markets, there have been structural changes in the global oil market. Financial instruments are communicating information about future conditions much more rapidly than in the past. Prices from long and short-duration contracts have started moving more together. Abrupt changes in supply and demand, influenced by such events and trends as the financial crisis of 2008-09, uncertainty about China’s economic growth rate, the Libyan uprising, the Iranian Nuclear standstill, and the Deepwater Horizon oil spill, change expectations and current prices. Although volatility appears greater over this period, financialization makes price discovery more robust. Most empirical economic studies suggest that fundamental factors shaped the expectations over 2004-08, although financial bubbles may have emerged just prior to and during the summer of 2008. This review represents a broad survey of economic research and literature related to the structure and functioning of the world oil market. The theories and models of oil demand and supply reviewed here, although imperfect in many respects, offer a clear and well-defined perspective on the forces that are shaping the markets for crude oil and refined products. Much work remains to be done if we are to achieve a more complete understanding of these forces and the trends that lie ahead. The contents that follow represent an assessment of how far we have come and where we are headed. Around the world governments, businesses and consumers share a vital interest in the benefits that flow from an efficient, well-functioning oil market. It is hoped, therefore, that the discussion in this review will find a broad audience.
The frequent fluctuations in oil price have caused great concern. Scholars have conducted a lot of researches on reasons for oil price fluctuations, but few examined the impact of the world economic cyclical changes on oil price cyclical fluctuations. This study applies wavelet analysis, Hodrik-Prescott Filter, Band-Pass Filter methods to measure oil price cyclical fluctuations, and applies vector autoregressive model to test correlation between economic cyclical changes and oil price cyclical fluctuations from two stages. The results show that the main cycle of oil price fluctuations is 6–7 years and there are two mutation points with an interval of 32 years, which reflects long-wave component of oil price fluctuations. In the first stage, there is a one-way Granger causality between oil price cyclical fluctuations and the world economic cyclical changes. In the second stage, there is a one-way Granger causality between the world economic cyclical changes and oil price cyclical fluctuations. When the world economic growth rate increases by 1%, oil price increases by 5.70%. The world economic cyclical changes contribute more to oil price cyclical fluctuations than OPEC's daily production. This study believes the world economic cyclical changes have a one-way causality to oil price cyclical fluctuations.